1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
2 <html><head><title>LLVM Programmer's Manual</title></head>
6 <table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
7 <tr><td> <font size=+3 color="#EEEEFF" face="Georgia,Palatino,Times,Roman"><b>LLVM Programmer's Manual</b></font></td>
11 <li><a href="#introduction">Introduction</a>
12 <li><a href="#general">General Information</a>
14 <li><a href="#stl">The C++ Standard Template Library</a>
16 <li>The <tt>-time-passes</tt> option
17 <li>How to use the LLVM Makefile system
18 <li>How to write a regression test
21 <li><a href="#apis">Important and useful LLVM APIs</a>
23 <li><a href="#isa">The <tt>isa<></tt>, <tt>cast<></tt> and
24 <tt>dyn_cast<></tt> templates</a>
25 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro &
26 <tt>-debug</tt> option</a>
28 <li><a href="#DEBUG_TYPE">Fine grained debug info with
29 <tt>DEBUG_TYPE</tt> and the <tt>-debug-only</tt> option</a/>
31 <li><a href="#Statistic">The <tt>Statistic</tt> template &
32 <tt>-stats</tt> option</a>
34 <li>The <tt>InstVisitor</tt> template
35 <li>The general graph API
38 <li><a href="#common">Helpful Hints for Common Operations</a>
40 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
42 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
43 in a <tt>Function</tt></a>
44 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
45 in a <tt>BasicBlock</tt></a>
46 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
47 in a <tt>Function</tt></a>
48 <li><a href="#iterate_convert">Turning an iterator into a class
50 <li><a href="#iterate_complex">Finding call sites: a more complex
52 <li><a href="#iterate_chains">Iterating over def-use & use-def
55 <li><a href="#simplechanges">Making simple changes</a>
57 <li><a href="#schanges_creating">Creating and inserting new
58 <tt>Instruction</tt>s</a>
59 <li><a href="#schanges_deleting">Deleting
60 <tt>Instruction</tt>s</a>
61 <li><a href="#schanges_replacing">Replacing an
62 <tt>Instruction</tt> with another <tt>Value</tt></a>
65 <li>Working with the Control Flow Graph
67 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
73 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
75 <li><a href="#Value">The <tt>Value</tt> class</a>
77 <li><a href="#User">The <tt>User</tt> class</a>
79 <li><a href="#Instruction">The <tt>Instruction</tt> class</a>
83 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
85 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a>
86 <li><a href="#Function">The <tt>Function</tt> class</a>
87 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a>
89 <li><a href="#Module">The <tt>Module</tt> class</a>
90 <li><a href="#Constant">The <tt>Constant</tt> class</a>
96 <li><a href="#Type">The <tt>Type</tt> class</a>
97 <li><a href="#Argument">The <tt>Argument</tt> class</a>
99 <li>The <tt>SymbolTable</tt> class
100 <li>The <tt>ilist</tt> and <tt>iplist</tt> classes
102 <li>Creating, inserting, moving and deleting from LLVM lists
104 <li>Important iterator invalidation semantics to be aware of
107 <p><b>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
108 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>, and
109 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a></b><p>
113 <!-- *********************************************************************** -->
114 <table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
115 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
116 <a name="introduction">Introduction
117 </b></font></td></tr></table><ul>
118 <!-- *********************************************************************** -->
120 This document is meant to highlight some of the important classes and interfaces
121 available in the LLVM source-base. This manual is not intended to explain what
122 LLVM is, how it works, and what LLVM code looks like. It assumes that you know
123 the basics of LLVM and are interested in writing transformations or otherwise
124 analyzing or manipulating the code.<p>
126 This document should get you oriented so that you can find your way in the
127 continuously growing source code that makes up the LLVM infrastructure. Note
128 that this manual is not intended to serve as a replacement for reading the
129 source code, so if you think there should be a method in one of these classes to
130 do something, but it's not listed, check the source. Links to the <a
131 href="/doxygen/">doxygen</a> sources are provided to make this as easy as
134 The first section of this document describes general information that is useful
135 to know when working in the LLVM infrastructure, and the second describes the
136 Core LLVM classes. In the future this manual will be extended with information
137 describing how to use extension libraries, such as dominator information, CFG
138 traversal routines, and useful utilities like the <tt><a
139 href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.<p>
142 <!-- *********************************************************************** -->
143 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
144 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
145 <a name="general">General Information
146 </b></font></td></tr></table><ul>
147 <!-- *********************************************************************** -->
149 This section contains general information that is useful if you are working in
150 the LLVM source-base, but that isn't specific to any particular API.<p>
153 <!-- ======================================================================= -->
154 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
155 <tr><td> </td><td width="100%">
156 <font color="#EEEEFF" face="Georgia,Palatino"><b>
157 <a name="stl">The C++ Standard Template Library</a>
158 </b></font></td></tr></table><ul>
160 LLVM makes heavy use of the C++ Standard Template Library (STL), perhaps much
161 more than you are used to, or have seen before. Because of this, you might want
162 to do a little background reading in the techniques used and capabilities of the
163 library. There are many good pages that discuss the STL, and several books on
164 the subject that you can get, so it will not be discussed in this document.<p>
166 Here are some useful links:<p>
168 <li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++
169 Library reference</a> - an excellent reference for the STL and other parts of
170 the standard C++ library.
172 <li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
173 O'Reilly book in the making. It has a decent <a
174 href="http://www.tempest-sw.com/cpp/ch13-libref.html">Standard Library
175 Reference</a> that rivals Dinkumware's, and is actually free until the book is
178 <li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
181 <li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
183 href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
186 <li><a href="http://www.research.att.com/~bs/C++.html">Bjarne Stroustrup's C++
191 You are also encouraged to take a look at the <a
192 href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
193 to write maintainable code more than where to put your curly braces.<p>
196 <!-- *********************************************************************** -->
197 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
198 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
199 <a name="apis">Important and useful LLVM APIs
200 </b></font></td></tr></table><ul>
201 <!-- *********************************************************************** -->
203 Here we highlight some LLVM APIs that are generally useful and good to know
204 about when writing transformations.<p>
206 <!-- ======================================================================= -->
207 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
208 <tr><td> </td><td width="100%">
209 <font color="#EEEEFF" face="Georgia,Palatino"><b>
210 <a name="isa">The isa<>, cast<> and dyn_cast<> templates</a>
211 </b></font></td></tr></table><ul>
213 The LLVM source-base makes extensive use of a custom form of RTTI. These
214 templates have many similarities to the C++ <tt>dynamic_cast<></tt>
215 operator, but they don't have some drawbacks (primarily stemming from the fact
216 that <tt>dynamic_cast<></tt> only works on classes that have a v-table).
217 Because they are used so often, you must know what they do and how they work.
218 All of these templates are defined in the <a
219 href="/doxygen/Casting_8h-source.html"><tt>Support/Casting.h</tt></a> file (note
220 that you very rarely have to include this file directly).<p>
224 <dt><tt>isa<></tt>:
226 <dd>The <tt>isa<></tt> operator works exactly like the Java
227 "<tt>instanceof</tt>" operator. It returns true or false depending on whether a
228 reference or pointer points to an instance of the specified class. This can be
229 very useful for constraint checking of various sorts (example below).<p>
232 <dt><tt>cast<></tt>:
234 <dd>The <tt>cast<></tt> operator is a "checked cast" operation. It
235 converts a pointer or reference from a base class to a derived cast, causing an
236 assertion failure if it is not really an instance of the right type. This
237 should be used in cases where you have some information that makes you believe
238 that something is of the right type. An example of the <tt>isa<></tt> and
239 <tt>cast<></tt> template is:<p>
242 static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
243 if (isa<<a href="#Constant">Constant</a>>(V) || isa<<a href="#Argument">Argument</a>>(V) || isa<<a href="#GlobalValue">GlobalValue</a>>(V))
246 <i>// Otherwise, it must be an instruction...</i>
247 return !L->contains(cast<<a href="#Instruction">Instruction</a>>(V)->getParent());
250 Note that you should <b>not</b> use an <tt>isa<></tt> test followed by a
251 <tt>cast<></tt>, for that use the <tt>dyn_cast<></tt> operator.<p>
254 <dt><tt>dyn_cast<></tt>:
256 <dd>The <tt>dyn_cast<></tt> operator is a "checking cast" operation. It
257 checks to see if the operand is of the specified type, and if so, returns a
258 pointer to it (this operator does not work with references). If the operand is
259 not of the correct type, a null pointer is returned. Thus, this works very much
260 like the <tt>dynamic_cast</tt> operator in C++, and should be used in the same
261 circumstances. Typically, the <tt>dyn_cast<></tt> operator is used in an
262 <tt>if</tt> statement or some other flow control statement like this:<p>
265 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast<<a href="#AllocationInst">AllocationInst</a>>(Val)) {
270 This form of the <tt>if</tt> statement effectively combines together a call to
271 <tt>isa<></tt> and a call to <tt>cast<></tt> into one statement,
272 which is very convenient.<p>
274 Another common example is:<p>
277 <i>// Loop over all of the phi nodes in a basic block</i>
278 BasicBlock::iterator BBI = BB->begin();
279 for (; <a href="#PhiNode">PHINode</a> *PN = dyn_cast<<a href="#PHINode">PHINode</a>>(BBI); ++BBI)
283 Note that the <tt>dyn_cast<></tt> operator, like C++'s
284 <tt>dynamic_cast</tt> or Java's <tt>instanceof</tt> operator, can be abused. In
285 particular you should not use big chained <tt>if/then/else</tt> blocks to check
286 for lots of different variants of classes. If you find yourself wanting to do
287 this, it is much cleaner and more efficient to use the InstVisitor class to
288 dispatch over the instruction type directly.<p>
291 <dt><tt>cast_or_null<></tt>:
293 <dd>The <tt>cast_or_null<></tt> operator works just like the
294 <tt>cast<></tt> operator, except that it allows for a null pointer as an
295 argument (which it then propagates). This can sometimes be useful, allowing you
296 to combine several null checks into one.<p>
299 <dt><tt>dyn_cast_or_null<></tt>:
301 <dd>The <tt>dyn_cast_or_null<></tt> operator works just like the
302 <tt>dyn_cast<></tt> operator, except that it allows for a null pointer as
303 an argument (which it then propagates). This can sometimes be useful, allowing
304 you to combine several null checks into one.<p>
308 These five templates can be used with any classes, whether they have a v-table
309 or not. To add support for these templates, you simply need to add
310 <tt>classof</tt> static methods to the class you are interested casting to.
311 Describing this is currently outside the scope of this document, but there are
312 lots of examples in the LLVM source base.<p>
315 <!-- ======================================================================= -->
316 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
317 <tr><td> </td><td width="100%">
318 <font color="#EEEEFF" face="Georgia,Palatino"><b>
319 <a name="DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt> option</a>
320 </b></font></td></tr></table><ul>
322 Often when working on your pass you will put a bunch of debugging printouts and
323 other code into your pass. After you get it working, you want to remove
324 it... but you may need it again in the future (to work out new bugs that you run
327 Naturally, because of this, you don't want to delete the debug printouts, but
328 you don't want them to always be noisy. A standard compromise is to comment
329 them out, allowing you to enable them if you need them in the future.<p>
331 The "<tt><a href="/doxygen/Debug_8h-source.html">Support/Debug.h</a></tt>" file
332 provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to this
333 problem. Basically, you can put arbitrary code into the argument of the
334 <tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
335 tool) is run with the '<tt>-debug</tt>' command line argument:
339 DEBUG(std::cerr << "I am here!\n");
343 Then you can run your pass like this:<p>
346 $ opt < a.bc > /dev/null -mypass
348 $ opt < a.bc > /dev/null -mypass -debug
353 Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you to
354 now have to create "yet another" command line option for the debug output for
355 your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
356 so they do not cause a performance impact at all (for the same reason, they
357 should also not contain side-effects!).<p>
359 One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
360 enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
361 "<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
362 program hasn't been started yet, you can always just run it with
365 <!-- _______________________________________________________________________ -->
366 </ul><h4><a name="DEBUG_TYPE"><hr size=0>Fine grained debug info with
367 <tt>DEBUG_TYPE()</tt> and the <tt>-debug-only</tt> option</a> </h4><ul>
369 Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
370 just turns on <b>too much</b> information (such as when working on the code
371 generator). If you want to enable debug information with more fine-grained
372 control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
373 option as follows:<p>
377 DEBUG(std::cerr << "No debug type\n");
379 #define DEBUG_TYPE "foo"
380 DEBUG(std::cerr << "'foo' debug type\n");
382 #define DEBUG_TYPE "bar"
383 DEBUG(std::cerr << "'bar' debug type\n");
385 #define DEBUG_TYPE ""
386 DEBUG(std::cerr << "No debug type (2)\n");
390 Then you can run your pass like this:<p>
393 $ opt < a.bc > /dev/null -mypass
395 $ opt < a.bc > /dev/null -mypass -debug
400 $ opt < a.bc > /dev/null -mypass -debug-only=foo
402 $ opt < a.bc > /dev/null -mypass -debug-only=bar
407 Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of a
408 file, to specify the debug type for the entire module (if you do this before you
409 <tt>#include "Support/Debug.h"</tt>, you don't have to insert the ugly
410 <tt>#undef</tt>'s). Also, you should use names more meaningful that "foo" and
411 "bar", because there is no system in place to ensure that names do not conflict:
412 if two different modules use the same string, they will all be turned on when
413 the name is specified. This allows all, say, instruction scheduling, debug
414 information to be enabled with <tt>-debug-type=InstrSched</tt>, even if the
415 source lives in multiple files.<p>
418 <!-- ======================================================================= -->
419 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
420 <tr><td> </td><td width="100%">
421 <font color="#EEEEFF" face="Georgia,Palatino"><b>
422 <a name="Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
424 </b></font></td></tr></table><ul>
427 href="/doxygen/Statistic_8h-source.html">Support/Statistic.h</a></tt>"
428 file provides a template named <tt>Statistic</tt> that is used as a unified way
429 to keeping track of what the LLVM compiler is doing and how effective various
430 optimizations are. It is useful to see what optimizations are contributing to
431 making a particular program run faster.<p>
433 Often you may run your pass on some big program, and you're interested to see
434 how many times it makes a certain transformation. Although you can do this with
435 hand inspection, or some ad-hoc method, this is a real pain and not very useful
436 for big programs. Using the <tt>Statistic</tt> template makes it very easy to
437 keep track of this information, and the calculated information is presented in a
438 uniform manner with the rest of the passes being executed.<p>
440 There are many examples of <tt>Statistic</tt> users, but this basics of using it
444 <li>Define your statistic like this:<p>
447 static Statistic<> NumXForms("mypassname", "The # of times I did stuff");
450 The <tt>Statistic</tt> template can emulate just about any data-type, but if you
451 do not specify a template argument, it defaults to acting like an unsigned int
452 counter (this is usually what you want).<p>
454 <li>Whenever you make a transformation, bump the counter:<p>
457 ++NumXForms; // I did stuff
462 That's all you have to do. To get '<tt>opt</tt>' to print out the statistics
463 gathered, use the '<tt>-stats</tt>' option:<p>
466 $ opt -stats -mypassname < program.bc > /dev/null
467 ... statistic output ...
470 When running <tt>gccas</tt> on a C file from the SPEC benchmark suite, it gives
471 a report that looks like this:<p>
474 7646 bytecodewriter - Number of normal instructions
475 725 bytecodewriter - Number of oversized instructions
476 129996 bytecodewriter - Number of bytecode bytes written
477 2817 raise - Number of insts DCEd or constprop'd
478 3213 raise - Number of cast-of-self removed
479 5046 raise - Number of expression trees converted
480 75 raise - Number of other getelementptr's formed
481 138 raise - Number of load/store peepholes
482 42 deadtypeelim - Number of unused typenames removed from symtab
483 392 funcresolve - Number of varargs functions resolved
484 27 globaldce - Number of global variables removed
485 2 adce - Number of basic blocks removed
486 134 cee - Number of branches revectored
487 49 cee - Number of setcc instruction eliminated
488 532 gcse - Number of loads removed
489 2919 gcse - Number of instructions removed
490 86 indvars - Number of cannonical indvars added
491 87 indvars - Number of aux indvars removed
492 25 instcombine - Number of dead inst eliminate
493 434 instcombine - Number of insts combined
494 248 licm - Number of load insts hoisted
495 1298 licm - Number of insts hoisted to a loop pre-header
496 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
497 75 mem2reg - Number of alloca's promoted
498 1444 cfgsimplify - Number of blocks simplified
501 Obviously, with so many optimizations, having a unified framework for this stuff
502 is very nice. Making your pass fit well into the framework makes it more
503 maintainable and useful.<p>
506 <!-- *********************************************************************** -->
507 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
508 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
509 <a name="common">Helpful Hints for Common Operations
510 </b></font></td></tr></table><ul> <!--
511 *********************************************************************** -->
513 This section describes how to perform some very simple transformations of LLVM
514 code. This is meant to give examples of common idioms used, showing the
515 practical side of LLVM transformations.<p>
517 Because this is a "how-to" section, you should also read about the main classes
518 that you will be working with. The <a href="#coreclasses">Core LLVM Class
519 Hierarchy Reference</a> contains details and descriptions of the main classes
520 that you should know about.<p>
522 <!-- NOTE: this section should be heavy on example code -->
525 <!-- ======================================================================= -->
526 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
527 <tr><td> </td><td width="100%">
528 <font color="#EEEEFF" face="Georgia,Palatino"><b>
529 <a name="inspection">Basic Inspection and Traversal Routines</a>
530 </b></font></td></tr></table><ul>
532 The LLVM compiler infrastructure have many different data structures that may be
533 traversed. Following the example of the C++ standard template library, the
534 techniques used to traverse these various data structures are all basically the
535 same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
536 method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
537 function returns an iterator pointing to one past the last valid element of the
538 sequence, and there is some <tt>XXXiterator</tt> data type that is common
539 between the two operations.<p>
541 Because the pattern for iteration is common across many different aspects of the
542 program representation, the standard template library algorithms may be used on
543 them, and it is easier to remember how to iterate. First we show a few common
544 examples of the data structures that need to be traversed. Other data
545 structures are traversed in very similar ways.<p>
548 <!-- _______________________________________________________________________ -->
549 </ul><h4><a name="iterate_function"><hr size=0>Iterating over the <a
550 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
551 href="#Function"><tt>Function</tt></a> </h4><ul>
553 It's quite common to have a <tt>Function</tt> instance that you'd like
554 to transform in some way; in particular, you'd like to manipulate its
555 <tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over
556 all of the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>.
557 The following is an example that prints the name of a
558 <tt>BasicBlock</tt> and the number of <tt>Instruction</tt>s it
562 // func is a pointer to a Function instance
563 for (Function::iterator i = func->begin(), e = func->end(); i != e; ++i) {
565 // print out the name of the basic block if it has one, and then the
566 // number of instructions that it contains
568 cerr << "Basic block (name=" << i->getName() << ") has "
569 << i->size() << " instructions.\n";
573 Note that i can be used as if it were a pointer for the purposes of
574 invoking member functions of the <tt>Instruction</tt> class. This is
575 because the indirection operator is overloaded for the iterator
576 classes. In the above code, the expression <tt>i->size()</tt> is
577 exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.
579 <!-- _______________________________________________________________________ -->
580 </ul><h4><a name="iterate_basicblock"><hr size=0>Iterating over the <a
581 href="#Instruction"><tt>Instruction</tt></a>s in a <a
582 href="#BasicBlock"><tt>BasicBlock</tt></a> </h4><ul>
584 Just like when dealing with <tt>BasicBlock</tt>s in
585 <tt>Function</tt>s, it's easy to iterate over the individual
586 instructions that make up <tt>BasicBlock</tt>s. Here's a code snippet
587 that prints out each instruction in a <tt>BasicBlock</tt>:
590 // blk is a pointer to a BasicBlock instance
591 for (BasicBlock::iterator i = blk->begin(), e = blk->end(); i != e; ++i)
592 // the next statement works since operator<<(ostream&,...)
593 // is overloaded for Instruction&
594 cerr << *i << "\n";
597 However, this isn't really the best way to print out the contents of a
598 <tt>BasicBlock</tt>! Since the ostream operators are overloaded for
599 virtually anything you'll care about, you could have just invoked the
600 print routine on the basic block itself: <tt>cerr << *blk <<
603 Note that currently operator<< is implemented for <tt>Value*</tt>, so it
604 will print out the contents of the pointer, instead of
605 the pointer value you might expect. This is a deprecated interface that will
606 be removed in the future, so it's best not to depend on it. To print out the
607 pointer value for now, you must cast to <tt>void*</tt>.<p>
610 <!-- _______________________________________________________________________ -->
611 </ul><h4><a name="iterate_institer"><hr size=0>Iterating over the <a
612 href="#Instruction"><tt>Instruction</tt></a>s in a <a
613 href="#Function"><tt>Function</tt></a></h4><ul>
615 If you're finding that you commonly iterate over a <tt>Function</tt>'s
616 <tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s
617 <tt>Instruction</tt>s, <tt>InstIterator</tt> should be used instead.
618 You'll need to include <a href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>, and then
619 instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
620 small example that shows how to dump all instructions in a function to
621 stderr (<b>Note:</b> Dereferencing an <tt>InstIterator</tt> yields an
622 <tt>Instruction*</tt>, <i>not</i> an <tt>Instruction&</tt>!):
625 #include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
627 // Suppose F is a ptr to a function
628 for (inst_iterator i = inst_begin(F), e = inst_end(F); i != e; ++i)
629 cerr << **i << "\n";
632 Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
633 worklist with its initial contents. For example, if you wanted to
634 initialize a worklist to contain all instructions in a
635 <tt>Function</tt> F, all you would need to do is something like:
638 std::set<Instruction*> worklist;
639 worklist.insert(inst_begin(F), inst_end(F));
642 The STL set <tt>worklist</tt> would now contain all instructions in
643 the <tt>Function</tt> pointed to by F.
645 <!-- _______________________________________________________________________ -->
646 </ul><h4><a name="iterate_convert"><hr size=0>Turning an iterator into a class
647 pointer (and vice-versa) </h4><ul>
649 Sometimes, it'll be useful to grab a reference (or pointer) to a class
650 instance when all you've got at hand is an iterator. Well, extracting
651 a reference or a pointer from an iterator is very straightforward.
652 Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and
653 <tt>j</tt> is a <tt>BasicBlock::const_iterator</tt>:
656 Instruction& inst = *i; // grab reference to instruction reference
657 Instruction* pinst = &*i; // grab pointer to instruction reference
658 const Instruction& inst = *j;
660 However, the iterators you'll be working with in the LLVM framework
661 are special: they will automatically convert to a ptr-to-instance type
662 whenever they need to. Instead of dereferencing the iterator and then
663 taking the address of the result, you can simply assign the iterator
664 to the proper pointer type and you get the dereference and address-of
665 operation as a result of the assignment (behind the scenes, this is a
666 result of overloading casting mechanisms). Thus the last line of the
669 <pre>Instruction* pinst = &*i;</pre>
671 is semantically equivalent to
673 <pre>Instruction* pinst = i;</pre>
675 It's also possible to turn a class pointer into the corresponding
676 iterator. Usually, this conversion is quite inexpensive. The
677 following code snippet illustrates use of the conversion constructors
678 provided by LLVM iterators. By using these, you can explicitly grab
679 the iterator of something without actually obtaining it via iteration
683 void printNextInstruction(Instruction* inst) {
684 BasicBlock::iterator it(inst);
685 ++it; // after this line, it refers to the instruction after *inst.
686 if (it != inst->getParent()->end()) cerr << *it << "\n";
689 Of course, this example is strictly pedagogical, because it'd be much
690 better to explicitly grab the next instruction directly from inst.
693 <!--_______________________________________________________________________-->
694 </ul><h4><a name="iterate_complex"><hr size=0>Finding call sites: a slightly
695 more complex example </h4><ul>
697 Say that you're writing a FunctionPass and would like to count all the
698 locations in the entire module (that is, across every
699 <tt>Function</tt>) where a certain function (i.e., some
700 <tt>Function</tt>*) is already in scope. As you'll learn later, you may
701 want to use an <tt>InstVisitor</tt> to accomplish this in a much more
702 straightforward manner, but this example will allow us to explore how
703 you'd do it if you didn't have <tt>InstVisitor</tt> around. In
704 pseudocode, this is what we want to do:
707 initialize callCounter to zero
708 for each Function f in the Module
709 for each BasicBlock b in f
710 for each Instruction i in b
711 if (i is a CallInst and calls the given function)
712 increment callCounter
715 And the actual code is (remember, since we're writing a
716 <tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply
717 has to override the <tt>runOnFunction</tt> method...):
720 Function* targetFunc = ...;
722 class OurFunctionPass : public FunctionPass {
724 OurFunctionPass(): callCounter(0) { }
726 virtual runOnFunction(Function& F) {
727 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
728 for (BasicBlock::iterator i = b->begin(); ie = b->end(); i != ie; ++i) {
729 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a><<a href="#CallInst">CallInst</a>>(&*i)) {
730 // we know we've encountered a call instruction, so we
731 // need to determine if it's a call to the
732 // function pointed to by m_func or not.
734 if (callInst->getCalledFunction() == targetFunc)
741 unsigned callCounter;
745 <!--_______________________________________________________________________-->
746 </ul><h4><a name="iterate_chains"><hr size=0>Iterating over def-use &
747 use-def chains</h4><ul>
749 Frequently, we might have an instance of the <a
750 href="/doxygen/classValue.html">Value Class</a> and we want to
751 determine which <tt>User</tt>s use the <tt>Value</tt>. The list of
752 all <tt>User</tt>s of a particular <tt>Value</tt> is called a
753 <i>def-use</i> chain. For example, let's say we have a
754 <tt>Function*</tt> named <tt>F</tt> to a particular function
755 <tt>foo</tt>. Finding all of the instructions that <i>use</i>
756 <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain of
762 for (Value::use_iterator i = F->use_begin(), e = F->use_end(); i != e; ++i) {
763 if (Instruction *Inst = dyn_cast<Instruction>(*i)) {
764 cerr << "F is used in instruction:\n";
765 cerr << *Inst << "\n";
770 Alternately, it's common to have an instance of the <a
771 href="/doxygen/classUser.html">User Class</a> and need to know what
772 <tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used
773 by a <tt>User</tt> is known as a <i>use-def</i> chain. Instances of
774 class <tt>Instruction</tt> are common <tt>User</tt>s, so we might want
775 to iterate over all of the values that a particular instruction uses
776 (that is, the operands of the particular <tt>Instruction</tt>):
779 Instruction* pi = ...;
781 for (User::op_iterator i = pi->op_begin(), e = pi->op_end(); i != e; ++i) {
789 def-use chains ("finding all users of"): Value::use_begin/use_end
790 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
793 <!-- ======================================================================= -->
794 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
795 <tr><td> </td><td width="100%">
796 <font color="#EEEEFF" face="Georgia,Palatino"><b>
797 <a name="simplechanges">Making simple changes</a>
798 </b></font></td></tr></table><ul>
800 There are some primitive transformation operations present in the LLVM
801 infrastructure that are worth knowing about. When performing
802 transformations, it's fairly common to manipulate the contents of
803 basic blocks. This section describes some of the common methods for
804 doing so and gives example code.
806 <!--_______________________________________________________________________-->
807 </ul><h4><a name="schanges_creating"><hr size=0>Creating and inserting
808 new <tt>Instruction</tt>s</h4><ul>
810 <i>Instantiating Instructions</i>
812 <p>Creation of <tt>Instruction</tt>s is straightforward: simply call the
813 constructor for the kind of instruction to instantiate and provide the
814 necessary parameters. For example, an <tt>AllocaInst</tt> only
815 <i>requires</i> a (const-ptr-to) <tt>Type</tt>. Thus:
817 <pre>AllocaInst* ai = new AllocaInst(Type::IntTy);</pre>
819 will create an <tt>AllocaInst</tt> instance that represents the
820 allocation of one integer in the current stack frame, at runtime.
821 Each <tt>Instruction</tt> subclass is likely to have varying default
822 parameters which change the semantics of the instruction, so refer to
823 the <a href="/doxygen/classInstruction.html">doxygen documentation for
824 the subclass of Instruction</a> that you're interested in
827 <p><i>Naming values</i></p>
830 It is very useful to name the values of instructions when you're able
831 to, as this facilitates the debugging of your transformations. If you
832 end up looking at generated LLVM machine code, you definitely want to
833 have logical names associated with the results of instructions! By
834 supplying a value for the <tt>Name</tt> (default) parameter of the
835 <tt>Instruction</tt> constructor, you associate a logical name with
836 the result of the instruction's execution at runtime. For example,
837 say that I'm writing a transformation that dynamically allocates space
838 for an integer on the stack, and that integer is going to be used as
839 some kind of index by some other code. To accomplish this, I place an
840 <tt>AllocaInst</tt> at the first point in the first
841 <tt>BasicBlock</tt> of some <tt>Function</tt>, and I'm intending to
842 use it within the same <tt>Function</tt>. I might do:
844 <pre>AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");</pre>
846 where <tt>indexLoc</tt> is now the logical name of the instruction's
847 execution value, which is a pointer to an integer on the runtime
851 <p><i>Inserting instructions</i></p>
854 There are essentially two ways to insert an <tt>Instruction</tt> into
855 an existing sequence of instructions that form a <tt>BasicBlock</tt>:
857 <li>Insertion into an explicit instruction list
859 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within
860 that <tt>BasicBlock</tt>, and a newly-created instruction
861 we wish to insert before <tt>*pi</tt>, we do the following:
864 BasicBlock *pb = ...;
865 Instruction *pi = ...;
866 Instruction *newInst = new Instruction(...);
867 pb->getInstList().insert(pi, newInst); // inserts newInst before pi in pb
871 <li>Insertion into an implicit instruction list
872 <p><tt>Instruction</tt> instances that are already in
873 <tt>BasicBlock</tt>s are implicitly associated with an existing
874 instruction list: the instruction list of the enclosing basic block.
875 Thus, we could have accomplished the same thing as the above code
876 without being given a <tt>BasicBlock</tt> by doing:
878 Instruction *pi = ...;
879 Instruction *newInst = new Instruction(...);
880 pi->getParent()->getInstList().insert(pi, newInst);
882 In fact, this sequence of steps occurs so frequently that the
883 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes
884 provide constructors which take (as a default parameter) a pointer to
885 an <tt>Instruction</tt> which the newly-created <tt>Instruction</tt>
886 should precede. That is, <tt>Instruction</tt> constructors are
887 capable of inserting the newly-created instance into the
888 <tt>BasicBlock</tt> of a provided instruction, immediately before that
889 instruction. Using an <tt>Instruction</tt> constructor with a
890 <tt>insertBefore</tt> (default) parameter, the above code becomes:
892 Instruction* pi = ...;
893 Instruction* newInst = new Instruction(..., pi);
895 which is much cleaner, especially if you're creating a lot of
896 instructions and adding them to <tt>BasicBlock</tt>s.
901 <!--_______________________________________________________________________-->
902 </ul><h4><a name="schanges_deleting"><hr size=0>Deleting
903 <tt>Instruction</tt>s</h4><ul>
905 Deleting an instruction from an existing sequence of instructions that form a <a
906 href="#BasicBlock"><tt>BasicBlock</tt></a> is very straightforward. First, you
907 must have a pointer to the instruction that you wish to delete. Second, you
908 need to obtain the pointer to that instruction's basic block. You use the
909 pointer to the basic block to get its list of instructions and then use the
910 erase function to remove your instruction.<p>
915 <a href="#Instruction">Instruction</a> *I = .. ;
916 <a href="#BasicBlock">BasicBlock</a> *BB = I->getParent();
917 BB->getInstList().erase(I);
920 <!--_______________________________________________________________________-->
921 </ul><h4><a name="schanges_replacing"><hr size=0>Replacing an
922 <tt>Instruction</tt> with another <tt>Value</tt></h4><ul>
924 <p><i>Replacing individual instructions</i></p>
927 href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>" permits use of two very useful replace functions:
928 <tt>ReplaceInstWithValue</tt> and <tt>ReplaceInstWithInst</tt>.
932 <li><tt>ReplaceInstWithValue</tt>
934 <p>This function replaces all uses (within a basic block) of a given
935 instruction with a value, and then removes the original instruction.
936 The following example illustrates the replacement of the result of a
937 particular <tt>AllocaInst</tt> that allocates memory for a single
938 integer with an null pointer to an integer.</p>
941 AllocaInst* instToReplace = ...;
942 BasicBlock::iterator ii(instToReplace);
943 ReplaceInstWithValue(instToReplace->getParent()->getInstList(), ii,
944 Constant::getNullValue(PointerType::get(Type::IntTy)));
947 <li><tt>ReplaceInstWithInst</tt>
949 <p>This function replaces a particular instruction with another
950 instruction. The following example illustrates the replacement of one
951 <tt>AllocaInst</tt> with another.<p>
954 AllocaInst* instToReplace = ...;
955 BasicBlock::iterator ii(instToReplace);
956 ReplaceInstWithInst(instToReplace->getParent()->getInstList(), ii,
957 new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));
961 <p><i>Replacing multiple uses of <tt>User</tt>s and
962 <tt>Value</tt>s</i></p>
964 You can use <tt>Value::replaceAllUsesWith</tt> and
965 <tt>User::replaceUsesOfWith</tt> to change more than one use at a
966 time. See the doxygen documentation for the <a
967 href="/doxygen/classValue.html">Value Class</a> and <a
968 href="/doxygen/classUser.html">User Class</a>, respectively, for more
971 <!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
972 include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
973 ReplaceInstWithValue, ReplaceInstWithInst
976 <!-- *********************************************************************** -->
977 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
978 <tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
979 <a name="coreclasses">The Core LLVM Class Hierarchy Reference
980 </b></font></td></tr></table><ul>
981 <!-- *********************************************************************** -->
983 The Core LLVM classes are the primary means of representing the program being
984 inspected or transformed. The core LLVM classes are defined in header files in
985 the <tt>include/llvm/</tt> directory, and implemented in the <tt>lib/VMCore</tt>
989 <!-- ======================================================================= -->
990 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
991 <tr><td> </td><td width="100%">
992 <font color="#EEEEFF" face="Georgia,Palatino"><b>
993 <a name="Value">The <tt>Value</tt> class</a>
994 </b></font></td></tr></table><ul>
996 <tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt></b><br>
997 doxygen info: <a href="/doxygen/classValue.html">Value Class</a><p>
1000 The <tt>Value</tt> class is the most important class in LLVM Source base. It
1001 represents a typed value that may be used (among other things) as an operand to
1002 an instruction. There are many different types of <tt>Value</tt>s, such as <a
1003 href="#Constant"><tt>Constant</tt></a>s, <a
1004 href="#Argument"><tt>Argument</tt></a>s, and even <a
1005 href="#Instruction"><tt>Instruction</tt></a>s and <a
1006 href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.<p>
1008 A particular <tt>Value</tt> may be used many times in the LLVM representation
1009 for a program. For example, an incoming argument to a function (represented
1010 with an instance of the <a href="#Argument">Argument</a> class) is "used" by
1011 every instruction in the function that references the argument. To keep track
1012 of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
1013 href="#User"><tt>User</tt></a>s that is using it (the <a
1014 href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
1015 graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
1016 def-use information in the program, and is accessible through the <tt>use_</tt>*
1017 methods, shown below.<p>
1019 Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed, and
1020 this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
1021 method. <a name="#nameWarning">In addition, all LLVM values can be named. The
1022 "name" of the <tt>Value</tt> is symbolic string printed in the LLVM code:<p>
1025 %<b>foo</b> = add int 1, 2
1028 The name of this instruction is "foo". <b>NOTE</b> that the name of any value
1029 may be missing (an empty string), so names should <b>ONLY</b> be used for
1030 debugging (making the source code easier to read, debugging printouts), they
1031 should not be used to keep track of values or map between them. For this
1032 purpose, use a <tt>std::map</tt> of pointers to the <tt>Value</tt> itself
1035 One important aspect of LLVM is that there is no distinction between an SSA
1036 variable and the operation that produces it. Because of this, any reference to
1037 the value produced by an instruction (or the value available as an incoming
1038 argument, for example) is represented as a direct pointer to the class that
1039 represents this value. Although this may take some getting used to, it
1040 simplifies the representation and makes it easier to manipulate.<p>
1043 <!-- _______________________________________________________________________ -->
1044 </ul><h4><a name="m_Value"><hr size=0>Important Public Members of
1045 the <tt>Value</tt> class</h4><ul>
1047 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the use-list<br>
1048 <tt>Value::use_const_iterator</tt>
1049 - Typedef for const_iterator over the use-list<br>
1050 <tt>unsigned use_size()</tt> - Returns the number of users of the value.<br>
1051 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
1052 <tt>use_iterator use_begin()</tt>
1053 - Get an iterator to the start of the use-list.<br>
1054 <tt>use_iterator use_end()</tt>
1055 - Get an iterator to the end of the use-list.<br>
1056 <tt><a href="#User">User</a> *use_back()</tt>
1057 - Returns the last element in the list.<p>
1059 These 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>
1061 <li><tt><a href="#Type">Type</a> *getType() const</tt><p>
1062 This method returns the Type of the Value.
1064 <li><tt>bool hasName() const</tt><br>
1065 <tt>std::string getName() const</tt><br>
1066 <tt>void setName(const std::string &Name)</tt><p>
1068 This family of methods is used to access and assign a name to a <tt>Value</tt>,
1069 be aware of the <a href="#nameWarning">precaution above</a>.<p>
1072 <li><tt>void replaceAllUsesWith(Value *V)</tt><p>
1074 This method traverses the use list of a <tt>Value</tt> changing all <a
1075 href="#User"><tt>User</tt>s</a> of the current value to refer to "<tt>V</tt>"
1076 instead. For example, if you detect that an instruction always produces a
1077 constant value (for example through constant folding), you can replace all uses
1078 of the instruction with the constant like this:<p>
1081 Inst->replaceAllUsesWith(ConstVal);
1086 <!-- ======================================================================= -->
1087 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1088 <tr><td> </td><td width="100%">
1089 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1090 <a name="User">The <tt>User</tt> class</a>
1091 </b></font></td></tr></table><ul>
1093 <tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt></b><br>
1094 doxygen info: <a href="/doxygen/classUser.html">User Class</a><br>
1095 Superclass: <a href="#Value"><tt>Value</tt></a><p>
1098 The <tt>User</tt> class is the common base class of all LLVM nodes that may
1099 refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
1100 that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
1101 referring to. The <tt>User</tt> class itself is a subclass of
1104 The operands of a <tt>User</tt> point directly to the LLVM <a
1105 href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
1106 Single Assignment (SSA) form, there can only be one definition referred to,
1107 allowing this direct connection. This connection provides the use-def
1108 information in LLVM.<p>
1110 <!-- _______________________________________________________________________ -->
1111 </ul><h4><a name="m_User"><hr size=0>Important Public Members of
1112 the <tt>User</tt> class</h4><ul>
1114 The <tt>User</tt> class exposes the operand list in two ways: through an index
1115 access interface and through an iterator based interface.<p>
1117 <li><tt>Value *getOperand(unsigned i)</tt><br>
1118 <tt>unsigned getNumOperands()</tt><p>
1120 These two methods expose the operands of the <tt>User</tt> in a convenient form
1121 for direct access.<p>
1123 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand list<br>
1124 <tt>User::op_const_iterator</tt>
1125 <tt>use_iterator op_begin()</tt>
1126 - Get an iterator to the start of the operand list.<br>
1127 <tt>use_iterator op_end()</tt>
1128 - Get an iterator to the end of the operand list.<p>
1130 Together, these methods make up the iterator based interface to the operands of
1135 <!-- ======================================================================= -->
1136 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1137 <tr><td> </td><td width="100%">
1138 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1139 <a name="Instruction">The <tt>Instruction</tt> class</a>
1140 </b></font></td></tr></table><ul>
1143 href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt></b><br>
1144 doxygen info: <a href="/doxygen/classInstruction.html">Instruction Class</a><br>
1145 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1146 href="#Value"><tt>Value</tt></a><p>
1148 The <tt>Instruction</tt> class is the common base class for all LLVM
1149 instructions. It provides only a few methods, but is a very commonly used
1150 class. The primary data tracked by the <tt>Instruction</tt> class itself is the
1151 opcode (instruction type) and the parent <a
1152 href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
1153 into. To represent a specific type of instruction, one of many subclasses of
1154 <tt>Instruction</tt> are used.<p>
1156 Because the <tt>Instruction</tt> class subclasses the <a
1157 href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
1158 way as for other <a href="#User"><tt>User</tt></a>s (with the
1159 <tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
1160 <tt>op_begin()</tt>/<tt>op_end()</tt> methods).<p>
1162 An important file for the <tt>Instruction</tt> class is the
1163 <tt>llvm/Instruction.def</tt> file. This file contains some meta-data about the
1164 various different types of instructions in LLVM. It describes the enum values
1165 that are used as opcodes (for example <tt>Instruction::Add</tt> and
1166 <tt>Instruction::SetLE</tt>), as well as the concrete sub-classes of
1167 <tt>Instruction</tt> that implement the instruction (for example <tt><a
1168 href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
1169 href="#SetCondInst">SetCondInst</a></tt>). Unfortunately, the use of macros in
1170 this file confused doxygen, so these enum values don't show up correctly in the
1171 <a href="/doxygen/classInstruction.html">doxygen output</a>.<p>
1174 <!-- _______________________________________________________________________ -->
1175 </ul><h4><a name="m_Instruction"><hr size=0>Important Public Members of
1176 the <tt>Instruction</tt> class</h4><ul>
1178 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt><p>
1180 Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that this
1181 <tt>Instruction</tt> is embedded into.<p>
1183 <li><tt>bool mayWriteToMemory()</tt><p>
1185 Returns true if the instruction writes to memory, i.e. it is a <tt>call</tt>,
1186 <tt>free</tt>, <tt>invoke</tt>, or <tt>store</tt>.<p>
1188 <li><tt>unsigned getOpcode()</tt><p>
1190 Returns the opcode for the <tt>Instruction</tt>.<p>
1192 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt><p>
1194 Returns another instance of the specified instruction, identical in all ways to
1195 the original except that the instruction has no parent (ie it's not embedded
1196 into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>), and it has no name.<p>
1202 \subsection{Subclasses of Instruction :}
1204 <li>BinaryOperator : This subclass of Instruction defines a general interface to the all the instructions involvong binary operators in LLVM.
1206 <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.
1208 <li>TerminatorInst : This subclass of Instructions defines an interface for all instructions that can terminate a BasicBlock.
1210 <li> <tt>unsigned getNumSuccessors()</tt>: Returns the number of successors for this terminator instruction.
1211 <li><tt>BasicBlock *getSuccessor(unsigned i)</tt>: As the name suggests returns the ith successor BasicBlock.
1212 <li><tt>void setSuccessor(unsigned i, BasicBlock *B)</tt>: sets BasicBlock B as the ith succesor to this terminator instruction.
1215 <li>PHINode : This represents the PHI instructions in the SSA form.
1217 <li><tt> unsigned getNumIncomingValues()</tt>: Returns the number of incoming edges to this PHI node.
1218 <li><tt> Value *getIncomingValue(unsigned i)</tt>: Returns the ith incoming Value.
1219 <li><tt>void setIncomingValue(unsigned i, Value *V)</tt>: Sets the ith incoming Value as V
1220 <li><tt>BasicBlock *getIncomingBlock(unsigned i)</tt>: Returns the Basic Block corresponding to the ith incoming Value.
1221 <li><tt> void addIncoming(Value *D, BasicBlock *BB)</tt>:
1222 Add an incoming value to the end of the PHI list
1223 <li><tt> int getBasicBlockIndex(const BasicBlock *BB) const</tt>:
1224 Returns the first index of the specified basic block in the value list for this PHI. Returns -1 if no instance.
1226 <li>CastInst : In LLVM all casts have to be done through explicit cast instructions. CastInst defines the interface to the cast instructions.
1227 <li>CallInst : This defines an interface to the call instruction in LLVM. ARguments to the function are nothing but operands of the instruction.
1229 <li>: <tt>Function *getCalledFunction()</tt>: Returns a handle to the function that is being called by this Function.
1231 <li>LoadInst, StoreInst, GetElemPtrInst : These subclasses represent load, store and getelementptr instructions in LLVM.
1233 <li><tt>Value * getPointerOperand()</tt>: Returns the Pointer Operand which is typically the 0th operand.
1235 <li>BranchInst : This is a subclass of TerminatorInst and defines the interface for conditional and unconditional branches in LLVM.
1237 <li><tt>bool isConditional()</tt>: Returns true if the branch is a conditional branch else returns false
1238 <li> <tt>Value *getCondition()</tt>: Returns the condition if it is a conditional branch else returns null.
1239 <li> <tt>void setUnconditionalDest(BasicBlock *Dest)</tt>: Changes the current branch to an unconditional one targetting the specified block.
1247 <!-- ======================================================================= -->
1248 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1249 <tr><td> </td><td width="100%">
1250 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1251 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
1252 </b></font></td></tr></table><ul>
1255 href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt></b><br>
1256 doxygen info: <a href="/doxygen/classBasicBlock.html">BasicBlock Class</a><br>
1257 Superclass: <a href="#Value"><tt>Value</tt></a><p>
1260 This class represents a single entry multiple exit section of the code, commonly
1261 known as a basic block by the compiler community. The <tt>BasicBlock</tt> class
1262 maintains a list of <a href="#Instruction"><tt>Instruction</tt></a>s, which form
1263 the body of the block. Matching the language definition, the last element of
1264 this list of instructions is always a terminator instruction (a subclass of the
1265 <a href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).<p>
1267 In addition to tracking the list of instructions that make up the block, the
1268 <tt>BasicBlock</tt> class also keeps track of the <a
1269 href="#Function"><tt>Function</tt></a> that it is embedded into.<p>
1271 Note that <tt>BasicBlock</tt>s themselves are <a
1272 href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
1273 like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
1277 <!-- _______________________________________________________________________ -->
1278 </ul><h4><a name="m_BasicBlock"><hr size=0>Important Public Members of
1279 the <tt>BasicBlock</tt> class</h4><ul>
1281 <li><tt>BasicBlock(const std::string &Name = "", <a
1282 href="#Function">Function</a> *Parent = 0)</tt><p>
1284 The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
1285 insertion into a function. The constructor simply takes a name for the new
1286 block, and optionally a <a href="#Function"><tt>Function</tt></a> to insert it
1287 into. If the <tt>Parent</tt> parameter is specified, the new
1288 <tt>BasicBlock</tt> is automatically inserted at the end of the specified <a
1289 href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
1290 manually inserted into the <a href="#Function"><tt>Function</tt></a>.<p>
1292 <li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
1293 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
1294 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1295 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt><p>
1297 These methods and typedefs are forwarding functions that have the same semantics
1298 as the standard library methods of the same names. These methods expose the
1299 underlying instruction list of a basic block in a way that is easy to
1300 manipulate. To get the full complement of container operations (including
1301 operations to update the list), you must use the <tt>getInstList()</tt>
1304 <li><tt>BasicBlock::InstListType &getInstList()</tt><p>
1306 This method is used to get access to the underlying container that actually
1307 holds the Instructions. This method must be used when there isn't a forwarding
1308 function in the <tt>BasicBlock</tt> class for the operation that you would like
1309 to perform. Because there are no forwarding functions for "updating"
1310 operations, you need to use this if you want to update the contents of a
1311 <tt>BasicBlock</tt>.<p>
1313 <li><tt><A href="#Function">Function</a> *getParent()</tt><p>
1315 Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
1316 embedded into, or a null pointer if it is homeless.<p>
1318 <li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt><p>
1320 Returns a pointer to the terminator instruction that appears at the end of the
1321 <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
1322 instruction in the block is not a terminator, then a null pointer is
1326 <!-- ======================================================================= -->
1327 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1328 <tr><td> </td><td width="100%">
1329 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1330 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
1331 </b></font></td></tr></table><ul>
1334 href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt></b><br>
1335 doxygen info: <a href="/doxygen/classGlobalValue.html">GlobalValue Class</a><br>
1336 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1337 href="#Value"><tt>Value</tt></a><p>
1339 Global values (<A href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
1340 href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
1341 visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
1342 Because they are visible at global scope, they are also subject to linking with
1343 other globals defined in different translation units. To control the linking
1344 process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
1345 <tt>GlobalValue</tt>s know whether they have internal or external linkage.<p>
1347 If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
1348 <tt>static</tt> in C), it is not visible to code outside the current translation
1349 unit, and does not participate in linking. If it has external linkage, it is
1350 visible to external code, and does participate in linking. In addition to
1351 linkage information, <tt>GlobalValue</tt>s keep track of which <a
1352 href="#Module"><tt>Module</tt></a> they are currently part of.<p>
1354 Because <tt>GlobalValue</tt>s are memory objects, they are always referred to by
1355 their address. As such, the <a href="#Type"><tt>Type</tt></a> of a global is
1356 always a pointer to its contents. This is explained in the LLVM Language
1357 Reference Manual.<p>
1360 <!-- _______________________________________________________________________ -->
1361 </ul><h4><a name="m_GlobalValue"><hr size=0>Important Public Members of
1362 the <tt>GlobalValue</tt> class</h4><ul>
1364 <li><tt>bool hasInternalLinkage() const</tt><br>
1365 <tt>bool hasExternalLinkage() const</tt><br>
1366 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt><p>
1368 These methods manipulate the linkage characteristics of the
1369 <tt>GlobalValue</tt>.<p>
1371 <li><tt><a href="#Module">Module</a> *getParent()</tt><p>
1373 This returns the <a href="#Module"><tt>Module</tt></a> that the GlobalValue is
1374 currently embedded into.<p>
1378 <!-- ======================================================================= -->
1379 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1380 <tr><td> </td><td width="100%">
1381 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1382 <a name="Function">The <tt>Function</tt> class</a>
1383 </b></font></td></tr></table><ul>
1386 href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt></b><br>
1387 doxygen info: <a href="/doxygen/classFunction.html">Function Class</a><br>
1388 Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1389 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a><p>
1391 The <tt>Function</tt> class represents a single procedure in LLVM. It is
1392 actually one of the more complex classes in the LLVM heirarchy because it must
1393 keep track of a large amount of data. The <tt>Function</tt> class keeps track
1394 of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal <a
1395 href="#Argument"><tt>Argument</tt></a>s, and a <a
1396 href="#SymbolTable"><tt>SymbolTable</tt></a>.<p>
1398 The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most commonly
1399 used part of <tt>Function</tt> objects. The list imposes an implicit ordering
1400 of the blocks in the function, which indicate how the code will be layed out by
1401 the backend. Additionally, the first <a
1402 href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
1403 <tt>Function</tt>. It is not legal in LLVM explicitly branch to this initial
1404 block. There are no implicit exit nodes, and in fact there may be multiple exit
1405 nodes from a single <tt>Function</tt>. If the <a
1406 href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
1407 the <tt>Function</tt> is actually a function declaration: the actual body of the
1408 function hasn't been linked in yet.<p>
1410 In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
1411 <tt>Function</tt> class also keeps track of the list of formal <a
1412 href="#Argument"><tt>Argument</tt></a>s that the function receives. This
1413 container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
1414 nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
1415 the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.<p>
1417 The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used LLVM
1418 feature that is only used when you have to look up a value by name. Aside from
1419 that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used internally to
1420 make sure that there are not conflicts between the names of <a
1421 href="#Instruction"><tt>Instruction</tt></a>s, <a
1422 href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
1423 href="#Argument"><tt>Argument</tt></a>s in the function body.<p>
1426 <!-- _______________________________________________________________________ -->
1427 </ul><h4><a name="m_Function"><hr size=0>Important Public Members of
1428 the <tt>Function</tt> class</h4><ul>
1430 <li><tt>Function(const <a href="#FunctionType">FunctionType</a> *Ty, bool isInternal, const std::string &N = "")</tt><p>
1432 Constructor used when you need to create new <tt>Function</tt>s to add the the
1433 program. The constructor must specify the type of the function to create and
1434 whether or not it should start out with internal or external linkage.<p>
1436 <li><tt>bool isExternal()</tt><p>
1438 Return whether or not the <tt>Function</tt> has a body defined. If the function
1439 is "external", it does not have a body, and thus must be resolved by linking
1440 with a function defined in a different translation unit.<p>
1443 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
1444 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
1445 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1446 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt><p>
1448 These are forwarding methods that make it easy to access the contents of a
1449 <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
1452 <li><tt>Function::BasicBlockListType &getBasicBlockList()</tt><p>
1454 Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This is
1455 neccesary to use when you need to update the list or perform a complex action
1456 that doesn't have a forwarding method.<p>
1459 <li><tt>Function::aiterator</tt> - Typedef for the argument list iterator<br>
1460 <tt>Function::const_aiterator</tt> - Typedef for const_iterator.<br>
1461 <tt>abegin()</tt>, <tt>aend()</tt>, <tt>afront()</tt>, <tt>aback()</tt>,
1462 <tt>asize()</tt>, <tt>aempty()</tt>, <tt>arbegin()</tt>, <tt>arend()</tt><p>
1464 These are forwarding methods that make it easy to access the contents of a
1465 <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a> list.<p>
1467 <li><tt>Function::ArgumentListType &getArgumentList()</tt><p>
1469 Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
1470 neccesary to use when you need to update the list or perform a complex action
1471 that doesn't have a forwarding method.<p>
1475 <li><tt><a href="#BasicBlock">BasicBlock</a> &getEntryNode()</tt><p>
1477 Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
1478 function. Because the entry block for the function is always the first block,
1479 this returns the first block of the <tt>Function</tt>.<p>
1481 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
1482 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt><p>
1484 This traverses the <a href="#Type"><tt>Type</tt></a> of the <tt>Function</tt>
1485 and returns the return type of the function, or the <a
1486 href="#FunctionType"><tt>FunctionType</tt></a> of the actual function.<p>
1488 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt><p>
1490 Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a> for this
1491 <tt>Function</tt>.<p>
1495 <!-- ======================================================================= -->
1496 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1497 <tr><td> </td><td width="100%">
1498 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1499 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
1500 </b></font></td></tr></table><ul>
1503 href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt></b><br>
1504 doxygen info: <a href="/doxygen/classGlobalVariable.html">GlobalVariable Class</a><br>
1505 Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1506 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a><p>
1508 Global variables are represented with the (suprise suprise)
1509 <tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are
1510 also subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such
1511 are always referenced by their address (global values must live in memory, so
1512 their "name" refers to their address). Global variables may have an initial
1513 value (which must be a <a href="#Constant"><tt>Constant</tt></a>), and if they
1514 have an initializer, they may be marked as "constant" themselves (indicating
1515 that their contents never change at runtime).<p>
1518 <!-- _______________________________________________________________________ -->
1519 </ul><h4><a name="m_GlobalVariable"><hr size=0>Important Public Members of the
1520 <tt>GlobalVariable</tt> class</h4><ul>
1522 <li><tt>GlobalVariable(const <a href="#Type">Type</a> *Ty, bool isConstant, bool
1523 isInternal, <a href="#Constant">Constant</a> *Initializer = 0, const std::string
1524 &Name = "")</tt><p>
1526 Create a new global variable of the specified type. If <tt>isConstant</tt> is
1527 true then the global variable will be marked as unchanging for the program, and
1528 if <tt>isInternal</tt> is true the resultant global variable will have internal
1529 linkage. Optionally an initializer and name may be specified for the global variable as well.<p>
1532 <li><tt>bool isConstant() const</tt><p>
1534 Returns true if this is a global variable is known not to be modified at
1538 <li><tt>bool hasInitializer()</tt><p>
1540 Returns true if this <tt>GlobalVariable</tt> has an intializer.<p>
1543 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt><p>
1545 Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal to call
1546 this method if there is no initializer.<p>
1549 <!-- ======================================================================= -->
1550 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1551 <tr><td> </td><td width="100%">
1552 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1553 <a name="Module">The <tt>Module</tt> class</a>
1554 </b></font></td></tr></table><ul>
1557 href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt></b><br>
1558 doxygen info: <a href="/doxygen/classModule.html">Module Class</a><p>
1560 The <tt>Module</tt> class represents the top level structure present in LLVM
1561 programs. An LLVM module is effectively either a translation unit of the
1562 original program or a combination of several translation units merged by the
1563 linker. The <tt>Module</tt> class keeps track of a list of <a
1564 href="#Function"><tt>Function</tt></a>s, a list of <a
1565 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
1566 href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
1567 helpful member functions that try to make common operations easy.<p>
1570 <!-- _______________________________________________________________________ -->
1571 </ul><h4><a name="m_Module"><hr size=0>Important Public Members of the
1572 <tt>Module</tt> class</h4><ul>
1574 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
1575 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
1576 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1577 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt><p>
1579 These are forwarding methods that make it easy to access the contents of a
1580 <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
1583 <li><tt>Module::FunctionListType &getFunctionList()</tt><p>
1585 Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
1586 neccesary to use when you need to update the list or perform a complex action
1587 that doesn't have a forwarding method.<p>
1589 <!-- Global Variable -->
1592 <li><tt>Module::giterator</tt> - Typedef for global variable list iterator<br>
1593 <tt>Module::const_giterator</tt> - Typedef for const_iterator.<br>
1594 <tt>gbegin()</tt>, <tt>gend()</tt>, <tt>gfront()</tt>, <tt>gback()</tt>,
1595 <tt>gsize()</tt>, <tt>gempty()</tt>, <tt>grbegin()</tt>, <tt>grend()</tt><p>
1597 These are forwarding methods that make it easy to access the contents of a
1598 <tt>Module</tt> object's <a href="#GlobalVariable"><tt>GlobalVariable</tt></a>
1601 <li><tt>Module::GlobalListType &getGlobalList()</tt><p>
1603 Returns the list of <a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s.
1604 This is neccesary to use when you need to update the list or perform a complex
1605 action that doesn't have a forwarding method.<p>
1608 <!-- Symbol table stuff -->
1611 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt><p>
1613 Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a> for
1614 this <tt>Module</tt>.<p>
1617 <!-- Convenience methods -->
1620 <li><tt><a href="#Function">Function</a> *getFunction(const std::string &Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt><p>
1622 Look up the specified function in the <tt>Module</tt> <a
1623 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
1627 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const std::string
1628 &Name, const <a href="#FunctionType">FunctionType</a> *T)</tt><p>
1630 Look up the specified function in the <tt>Module</tt> <a
1631 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
1632 external declaration for the function and return it.<p>
1635 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt><p>
1637 If there is at least one entry in the <a
1638 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
1639 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
1643 <li><tt>bool addTypeName(const std::string &Name, const <a href="#Type">Type</a>
1646 Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a> mapping
1647 <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this name, true
1648 is returned and the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is not
1652 <!-- ======================================================================= -->
1653 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1654 <tr><td> </td><td width="100%">
1655 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1656 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
1657 </b></font></td></tr></table><ul>
1659 Constant represents a base class for different types of constants. It is
1660 subclassed by ConstantBool, ConstantInt, ConstantSInt, ConstantUInt,
1661 ConstantArray etc for representing the various types of Constants.<p>
1664 <!-- _______________________________________________________________________ -->
1665 </ul><h4><a name="m_Value"><hr size=0>Important Public Methods</h4><ul>
1667 <li><tt>bool isConstantExpr()</tt>: Returns true if it is a ConstantExpr
1671 Important Subclasses of Constant<p>
1674 <li>ConstantSInt : This subclass of Constant represents a signed integer constant.
1676 <li><tt>int64_t getValue() const</tt>: Returns the underlying value of this constant.
1678 <li>ConstantUInt : This class represents an unsigned integer.
1680 <li><tt>uint64_t getValue() const</tt>: Returns the underlying value of this constant.
1682 <li>ConstantFP : This class represents a floating point constant.
1684 <li><tt>double getValue() const</tt>: Returns the underlying value of this constant.
1686 <li>ConstantBool : This represents a boolean constant.
1688 <li><tt>bool getValue() const</tt>: Returns the underlying value of this constant.
1690 <li>ConstantArray : This represents a constant array.
1692 <li><tt>const std::vector<Use> &getValues() const</tt>: Returns a Vecotr of component constants that makeup this array.
1694 <li>ConstantStruct : This represents a constant struct.
1696 <li><tt>const std::vector<Use> &getValues() const</tt>: Returns a Vecotr of component constants that makeup this array.
1698 <li>ConstantPointerRef : This represents a constant pointer value that is initialized to point to a global value, which lies at a constant fixed address.
1700 <li><tt>GlobalValue *getValue()</tt>: Returns the global value to which this pointer is pointing to.
1705 <!-- ======================================================================= -->
1706 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1707 <tr><td> </td><td width="100%">
1708 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1709 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
1710 </b></font></td></tr></table><ul>
1712 Type as noted earlier is also a subclass of a Value class. Any primitive
1713 type (like int, short etc) in LLVM is an instance of Type Class. All
1714 other types are instances of subclasses of type like FunctionType,
1715 ArrayType etc. DerivedType is the interface for all such dervied types
1716 including FunctionType, ArrayType, PointerType, StructType. Types can have
1717 names. They can be recursive (StructType). There exists exactly one instance
1718 of any type structure at a time. This allows using pointer equality of Type *s for comparing types.
1720 <!-- _______________________________________________________________________ -->
1721 </ul><h4><a name="m_Value"><hr size=0>Important Public Methods</h4><ul>
1723 <li><tt>PrimitiveID getPrimitiveID() const</tt>: Returns the base type of the type.
1724 <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.
1725 <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.
1726 <li><tt> bool isInteger() const</tt>: Equilivent to isSigned() || isUnsigned(), but with only a single virtual function invocation.
1727 <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.
1729 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two floating point types.
1730 <li><tt>bool isRecursive() const</tt>: Returns rue if the type graph contains a cycle.
1731 <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.
1732 <li><tt>bool isPrimitiveType() const</tt>: Returns true if it is a primitive type.
1733 <li><tt>bool isDerivedType() const</tt>: Returns true if it is a derived type.
1734 <li><tt>const Type * getContainedType (unsigned i) const</tt>:
1735 This 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.
1736 <li><tt>unsigned getNumContainedTypes() const</tt>: Return the number of types in the derived type.
1744 <li>SequentialType : This is subclassed by ArrayType and PointerType
1746 <li><tt>const Type * getElementType() const</tt>: Returns the type of each of the elements in the sequential type.
1748 <li>ArrayType : This is a subclass of SequentialType and defines interface for array types.
1750 <li><tt>unsigned getNumElements() const</tt>: Returns the number of elements in the array.
1752 <li>PointerType : Subclass of SequentialType for pointer types.
1753 <li>StructType : subclass of DerivedTypes for struct types
1754 <li>FunctionType : subclass of DerivedTypes for function types.
1758 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg function
1759 <li><tt> const Type * getReturnType() const</tt>: Returns the return type of the function.
1760 <li><tt> const ParamTypes &getParamTypes() const</tt>: Returns a vector of parameter types.
1761 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns the type of the ith parameter.
1762 <li><tt> const unsigned getNumParams() const</tt>: Returns the number of formal parameters.
1769 <!-- ======================================================================= -->
1770 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0>
1771 <tr><td> </td><td width="100%">
1772 <font color="#EEEEFF" face="Georgia,Palatino"><b>
1773 <a name="Argument">The <tt>Argument</tt> class</a>
1774 </b></font></td></tr></table><ul>
1776 This subclass of Value defines the interface for incoming formal arguments to a
1777 function. A Function maitanis a list of its formal arguments. An argument has a
1778 pointer to the parent Function.
1783 <!-- *********************************************************************** -->
1785 <!-- *********************************************************************** -->
1788 <address>By: <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
1789 <a href="mailto:sabre@nondot.org">Chris Lattner</a></address>
1790 <!-- Created: Tue Aug 6 15:00:33 CDT 2002 -->
1791 <!-- hhmts start -->
1792 Last modified: Tue Aug 5 17:53:43 CDT 2003
1794 </font></body></html>