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></li>
20 <li>The <tt>-time-passes</tt> option</li>
21 <li>How to use the LLVM Makefile system</li>
22 <li>How to write a regression test</li>
27 <li><a href="#apis">Important and useful LLVM APIs</a>
29 <li><a href="#isa">The <tt>isa<></tt>, <tt>cast<></tt>
30 and <tt>dyn_cast<></tt> templates</a> </li>
31 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt>
34 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
35 and the <tt>-debug-only</tt> option</a> </li>
38 <li><a href="#Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
41 <li>The <tt>InstVisitor</tt> template
42 <li>The general graph API
44 <li><a href="#ViewGraph">Viewing graphs while debugging code</a></li>
47 <li><a href="#common">Helpful Hints for Common Operations</a>
49 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
51 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
52 in a <tt>Function</tt></a> </li>
53 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
54 in a <tt>BasicBlock</tt></a> </li>
55 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
56 in a <tt>Function</tt></a> </li>
57 <li><a href="#iterate_convert">Turning an iterator into a
58 class pointer</a> </li>
59 <li><a href="#iterate_complex">Finding call sites: a more
60 complex example</a> </li>
61 <li><a href="#calls_and_invokes">Treating calls and invokes
62 the same way</a> </li>
63 <li><a href="#iterate_chains">Iterating over def-use &
64 use-def chains</a> </li>
67 <li><a href="#simplechanges">Making simple changes</a>
69 <li><a href="#schanges_creating">Creating and inserting new
70 <tt>Instruction</tt>s</a> </li>
71 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
72 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
73 with another <tt>Value</tt></a> </li>
77 <li>Working with the Control Flow Graph
79 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
87 <li><a href="#advanced">Advanced Topics</a>
89 <li><a href="#TypeResolve">LLVM Type Resolution</a>
91 <li><a href="#BuildRecType">Basic Recursive Type Construction</a></li>
92 <li><a href="#refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a></li>
93 <li><a href="#PATypeHolder">The PATypeHolder Class</a></li>
94 <li><a href="#AbstractTypeUser">The AbstractTypeUser Class</a></li>
97 <li><a href="#SymbolTable">The <tt>SymbolTable</tt> class </a></li>
100 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
102 <li><a href="#Value">The <tt>Value</tt> class</a>
104 <li><a href="#User">The <tt>User</tt> class</a>
106 <li><a href="#Instruction">The <tt>Instruction</tt> class</a>
108 <li><a href="#GetElementPtrInst">The <tt>GetElementPtrInst</tt> class</a></li>
111 <li><a href="#Module">The <tt>Module</tt> class</a></li>
112 <li><a href="#Constant">The <tt>Constant</tt> class</a>
114 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
116 <li><a href="#BasicBlock">The <tt>BasicBlock</tt>class</a></li>
117 <li><a href="#Function">The <tt>Function</tt> class</a></li>
118 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
125 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
126 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
133 <div class="doc_author">
134 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
135 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
136 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>, and
137 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a></p>
140 <!-- *********************************************************************** -->
141 <div class="doc_section">
142 <a name="introduction">Introduction </a>
144 <!-- *********************************************************************** -->
146 <div class="doc_text">
148 <p>This document is meant to highlight some of the important classes and
149 interfaces available in the LLVM source-base. This manual is not
150 intended to explain what LLVM is, how it works, and what LLVM code looks
151 like. It assumes that you know the basics of LLVM and are interested
152 in writing transformations or otherwise analyzing or manipulating the
155 <p>This document should get you oriented so that you can find your
156 way in the continuously growing source code that makes up the LLVM
157 infrastructure. Note that this manual is not intended to serve as a
158 replacement for reading the source code, so if you think there should be
159 a method in one of these classes to do something, but it's not listed,
160 check the source. Links to the <a href="/doxygen/">doxygen</a> sources
161 are provided to make this as easy as possible.</p>
163 <p>The first section of this document describes general information that is
164 useful to know when working in the LLVM infrastructure, and the second describes
165 the Core LLVM classes. In the future this manual will be extended with
166 information describing how to use extension libraries, such as dominator
167 information, CFG traversal routines, and useful utilities like the <tt><a
168 href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
172 <!-- *********************************************************************** -->
173 <div class="doc_section">
174 <a name="general">General Information</a>
176 <!-- *********************************************************************** -->
178 <div class="doc_text">
180 <p>This section contains general information that is useful if you are working
181 in the LLVM source-base, but that isn't specific to any particular API.</p>
185 <!-- ======================================================================= -->
186 <div class="doc_subsection">
187 <a name="stl">The C++ Standard Template Library</a>
190 <div class="doc_text">
192 <p>LLVM makes heavy use of the C++ Standard Template Library (STL),
193 perhaps much more than you are used to, or have seen before. Because of
194 this, you might want to do a little background reading in the
195 techniques used and capabilities of the library. There are many good
196 pages that discuss the STL, and several books on the subject that you
197 can get, so it will not be discussed in this document.</p>
199 <p>Here are some useful links:</p>
203 <li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
204 reference</a> - an excellent reference for the STL and other parts of the
205 standard C++ library.</li>
207 <li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
208 O'Reilly book in the making. It has a decent
210 Reference that rivals Dinkumware's, and is unfortunately no longer free since the book has been
213 <li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
216 <li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
218 href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
221 <li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
224 <li><a href="http://64.78.49.204/">
225 Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
230 <p>You are also encouraged to take a look at the <a
231 href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
232 to write maintainable code more than where to put your curly braces.</p>
236 <!-- ======================================================================= -->
237 <div class="doc_subsection">
238 <a name="stl">Other useful references</a>
241 <div class="doc_text">
244 <li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
245 Branch and Tag Primer</a></li>
246 <li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
247 static and shared libraries across platforms</a></li>
252 <!-- *********************************************************************** -->
253 <div class="doc_section">
254 <a name="apis">Important and useful LLVM APIs</a>
256 <!-- *********************************************************************** -->
258 <div class="doc_text">
260 <p>Here we highlight some LLVM APIs that are generally useful and good to
261 know about when writing transformations.</p>
265 <!-- ======================================================================= -->
266 <div class="doc_subsection">
267 <a name="isa">The isa<>, cast<> and dyn_cast<> templates</a>
270 <div class="doc_text">
272 <p>The LLVM source-base makes extensive use of a custom form of RTTI.
273 These templates have many similarities to the C++ <tt>dynamic_cast<></tt>
274 operator, but they don't have some drawbacks (primarily stemming from
275 the fact that <tt>dynamic_cast<></tt> only works on classes that
276 have a v-table). Because they are used so often, you must know what they
277 do and how they work. All of these templates are defined in the <a
278 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
279 file (note that you very rarely have to include this file directly).</p>
282 <dt><tt>isa<></tt>: </dt>
284 <dd>The <tt>isa<></tt> operator works exactly like the Java
285 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
286 a reference or pointer points to an instance of the specified class. This can
287 be very useful for constraint checking of various sorts (example below).</dd>
289 <dt><tt>cast<></tt>: </dt>
291 <dd>The <tt>cast<></tt> operator is a "checked cast" operation. It
292 converts a pointer or reference from a base class to a derived cast, causing
293 an assertion failure if it is not really an instance of the right type. This
294 should be used in cases where you have some information that makes you believe
295 that something is of the right type. An example of the <tt>isa<></tt>
296 and <tt>cast<></tt> template is:
299 static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
300 if (isa<<a href="#Constant">Constant</a>>(V) || isa<<a href="#Argument">Argument</a>>(V) || isa<<a href="#GlobalValue">GlobalValue</a>>(V))
303 <i>// Otherwise, it must be an instruction...</i>
304 return !L->contains(cast<<a href="#Instruction">Instruction</a>>(V)->getParent());
308 <p>Note that you should <b>not</b> use an <tt>isa<></tt> test followed
309 by a <tt>cast<></tt>, for that use the <tt>dyn_cast<></tt>
314 <dt><tt>dyn_cast<></tt>:</dt>
316 <dd>The <tt>dyn_cast<></tt> operator is a "checking cast" operation. It
317 checks to see if the operand is of the specified type, and if so, returns a
318 pointer to it (this operator does not work with references). If the operand is
319 not of the correct type, a null pointer is returned. Thus, this works very
320 much like the <tt>dynamic_cast</tt> operator in C++, and should be used in the
321 same circumstances. Typically, the <tt>dyn_cast<></tt> operator is used
322 in an <tt>if</tt> statement or some other flow control statement like this:
325 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast<<a href="#AllocationInst">AllocationInst</a>>(Val)) {
330 <p> This form of the <tt>if</tt> statement effectively combines together a
331 call to <tt>isa<></tt> and a call to <tt>cast<></tt> into one
332 statement, which is very convenient.</p>
334 <p>Note that the <tt>dyn_cast<></tt> operator, like C++'s
335 <tt>dynamic_cast</tt> or Java's <tt>instanceof</tt> operator, can be abused.
336 In particular you should not use big chained <tt>if/then/else</tt> blocks to
337 check for lots of different variants of classes. If you find yourself
338 wanting to do this, it is much cleaner and more efficient to use the
339 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
343 <dt><tt>cast_or_null<></tt>: </dt>
345 <dd>The <tt>cast_or_null<></tt> operator works just like the
346 <tt>cast<></tt> operator, except that it allows for a null pointer as
347 an argument (which it then propagates). This can sometimes be useful,
348 allowing you to combine several null checks into one.</dd>
350 <dt><tt>dyn_cast_or_null<></tt>: </dt>
352 <dd>The <tt>dyn_cast_or_null<></tt> operator works just like the
353 <tt>dyn_cast<></tt> operator, except that it allows for a null pointer
354 as an argument (which it then propagates). This can sometimes be useful,
355 allowing you to combine several null checks into one.</dd>
359 <p>These five templates can be used with any classes, whether they have a
360 v-table or not. To add support for these templates, you simply need to add
361 <tt>classof</tt> static methods to the class you are interested casting
362 to. Describing this is currently outside the scope of this document, but there
363 are lots of examples in the LLVM source base.</p>
367 <!-- ======================================================================= -->
368 <div class="doc_subsection">
369 <a name="DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt> option</a>
372 <div class="doc_text">
374 <p>Often when working on your pass you will put a bunch of debugging printouts
375 and other code into your pass. After you get it working, you want to remove
376 it... but you may need it again in the future (to work out new bugs that you run
379 <p> Naturally, because of this, you don't want to delete the debug printouts,
380 but you don't want them to always be noisy. A standard compromise is to comment
381 them out, allowing you to enable them if you need them in the future.</p>
383 <p>The "<tt><a href="/doxygen/Debug_8h-source.html">llvm/Support/Debug.h</a></tt>"
384 file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
385 this problem. Basically, you can put arbitrary code into the argument of the
386 <tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
387 tool) is run with the '<tt>-debug</tt>' command line argument:</p>
389 <pre> ... <br> DEBUG(std::cerr << "I am here!\n");<br> ...<br></pre>
391 <p>Then you can run your pass like this:</p>
393 <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>
395 <p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
396 to not have to create "yet another" command line option for the debug output for
397 your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
398 so they do not cause a performance impact at all (for the same reason, they
399 should also not contain side-effects!).</p>
401 <p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
402 enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
403 "<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
404 program hasn't been started yet, you can always just run it with
409 <!-- _______________________________________________________________________ -->
410 <div class="doc_subsubsection">
411 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
412 the <tt>-debug-only</tt> option</a>
415 <div class="doc_text">
417 <p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
418 just turns on <b>too much</b> information (such as when working on the code
419 generator). If you want to enable debug information with more fine-grained
420 control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
421 option as follows:</p>
423 <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>
425 <p>Then you can run your pass like this:</p>
427 <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>
429 <p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
430 a file, to specify the debug type for the entire module (if you do this before
431 you <tt>#include "llvm/Support/Debug.h"</tt>, you don't have to insert the ugly
432 <tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
433 "bar", because there is no system in place to ensure that names do not
434 conflict. If two different modules use the same string, they will all be turned
435 on when the name is specified. This allows, for example, all debug information
436 for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
437 even if the source lives in multiple files.</p>
441 <!-- ======================================================================= -->
442 <div class="doc_subsection">
443 <a name="Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
447 <div class="doc_text">
450 href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
451 provides a template named <tt>Statistic</tt> that is used as a unified way to
452 keep track of what the LLVM compiler is doing and how effective various
453 optimizations are. It is useful to see what optimizations are contributing to
454 making a particular program run faster.</p>
456 <p>Often you may run your pass on some big program, and you're interested to see
457 how many times it makes a certain transformation. Although you can do this with
458 hand inspection, or some ad-hoc method, this is a real pain and not very useful
459 for big programs. Using the <tt>Statistic</tt> template makes it very easy to
460 keep track of this information, and the calculated information is presented in a
461 uniform manner with the rest of the passes being executed.</p>
463 <p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
464 it are as follows:</p>
467 <li>Define your statistic like this:
468 <pre>static Statistic<> NumXForms("mypassname", "The # of times I did stuff");<br></pre>
470 <p>The <tt>Statistic</tt> template can emulate just about any data-type,
471 but if you do not specify a template argument, it defaults to acting like
472 an unsigned int counter (this is usually what you want).</p></li>
474 <li>Whenever you make a transformation, bump the counter:
475 <pre> ++NumXForms; // I did stuff<br></pre>
479 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
480 statistics gathered, use the '<tt>-stats</tt>' option:</p>
482 <pre> $ opt -stats -mypassname < program.bc > /dev/null<br> ... statistic output ...<br></pre>
484 <p> When running <tt>gccas</tt> on a C file from the SPEC benchmark
485 suite, it gives a report that looks like this:</p>
487 <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>
489 <p>Obviously, with so many optimizations, having a unified framework for this
490 stuff is very nice. Making your pass fit well into the framework makes it more
491 maintainable and useful.</p>
495 <!-- ======================================================================= -->
496 <div class="doc_subsection">
497 <a name="ViewGraph">Viewing graphs while debugging code</a>
500 <div class="doc_text">
502 <p>Several of the important data structures in LLVM are graphs: for example
503 CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
504 LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
505 <a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
506 DAGs</a>. In many cases, while debugging various parts of the compiler, it is
507 nice to instantly visualize these graphs.</p>
509 <p>LLVM provides several callbacks that are available in a debug build to do
510 exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
511 the current LLVM tool will pop up a window containing the CFG for the function
512 where each basic block is a node in the graph, and each node contains the
513 instructions in the block. Similarly, there also exists
514 <tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
515 <tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
516 and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
517 you can usually use something like "<tt>call DAG.viewGraph()</tt>" to pop
518 up a window. Alternatively, you can sprinkle calls to these functions in your
519 code in places you want to debug.</p>
521 <p>Getting this to work requires a small amount of configuration. On Unix
522 systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
523 toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
524 Mac OS/X, download and install the Mac OS/X <a
525 href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
526 <tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or whereever you install
527 it) to your path. Once in your system and path are set up, rerun the LLVM
528 configure script and rebuild LLVM to enable this functionality.</p>
533 <!-- *********************************************************************** -->
534 <div class="doc_section">
535 <a name="common">Helpful Hints for Common Operations</a>
537 <!-- *********************************************************************** -->
539 <div class="doc_text">
541 <p>This section describes how to perform some very simple transformations of
542 LLVM code. This is meant to give examples of common idioms used, showing the
543 practical side of LLVM transformations. <p> Because this is a "how-to" section,
544 you should also read about the main classes that you will be working with. The
545 <a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
546 and descriptions of the main classes that you should know about.</p>
550 <!-- NOTE: this section should be heavy on example code -->
551 <!-- ======================================================================= -->
552 <div class="doc_subsection">
553 <a name="inspection">Basic Inspection and Traversal Routines</a>
556 <div class="doc_text">
558 <p>The LLVM compiler infrastructure have many different data structures that may
559 be traversed. Following the example of the C++ standard template library, the
560 techniques used to traverse these various data structures are all basically the
561 same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
562 method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
563 function returns an iterator pointing to one past the last valid element of the
564 sequence, and there is some <tt>XXXiterator</tt> data type that is common
565 between the two operations.</p>
567 <p>Because the pattern for iteration is common across many different aspects of
568 the program representation, the standard template library algorithms may be used
569 on them, and it is easier to remember how to iterate. First we show a few common
570 examples of the data structures that need to be traversed. Other data
571 structures are traversed in very similar ways.</p>
575 <!-- _______________________________________________________________________ -->
576 <div class="doc_subsubsection">
577 <a name="iterate_function">Iterating over the </a><a
578 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
579 href="#Function"><tt>Function</tt></a>
582 <div class="doc_text">
584 <p>It's quite common to have a <tt>Function</tt> instance that you'd like to
585 transform in some way; in particular, you'd like to manipulate its
586 <tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
587 the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
588 an example that prints the name of a <tt>BasicBlock</tt> and the number of
589 <tt>Instruction</tt>s it contains:</p>
591 <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>
593 <p>Note that i can be used as if it were a pointer for the purposes of
594 invoking member functions of the <tt>Instruction</tt> class. This is
595 because the indirection operator is overloaded for the iterator
596 classes. In the above code, the expression <tt>i->size()</tt> is
597 exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
601 <!-- _______________________________________________________________________ -->
602 <div class="doc_subsubsection">
603 <a name="iterate_basicblock">Iterating over the </a><a
604 href="#Instruction"><tt>Instruction</tt></a>s in a <a
605 href="#BasicBlock"><tt>BasicBlock</tt></a>
608 <div class="doc_text">
610 <p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
611 easy to iterate over the individual instructions that make up
612 <tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
613 a <tt>BasicBlock</tt>:</p>
616 // blk is a pointer to a BasicBlock instance
617 for (BasicBlock::iterator i = blk->begin(), e = blk->end(); i != e; ++i)
618 // the next statement works since operator<<(ostream&,...)
619 // is overloaded for Instruction&
620 std::cerr << *i << "\n";
623 <p>However, this isn't really the best way to print out the contents of a
624 <tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
625 anything you'll care about, you could have just invoked the print routine on the
626 basic block itself: <tt>std::cerr << *blk << "\n";</tt>.</p>
630 <!-- _______________________________________________________________________ -->
631 <div class="doc_subsubsection">
632 <a name="iterate_institer">Iterating over the </a><a
633 href="#Instruction"><tt>Instruction</tt></a>s in a <a
634 href="#Function"><tt>Function</tt></a>
637 <div class="doc_text">
639 <p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
640 <tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
641 <tt>InstIterator</tt> should be used instead. You'll need to include <a
642 href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
643 and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
644 small example that shows how to dump all instructions in a function to the standard error stream:<p>
646 <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>
647 Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
648 worklist with its initial contents. For example, if you wanted to
649 initialize a worklist to contain all instructions in a <tt>Function</tt>
650 F, all you would need to do is something like:
651 <pre>std::set<Instruction*> worklist;<br>worklist.insert(inst_begin(F), inst_end(F));<br></pre>
653 <p>The STL set <tt>worklist</tt> would now contain all instructions in the
654 <tt>Function</tt> pointed to by F.</p>
658 <!-- _______________________________________________________________________ -->
659 <div class="doc_subsubsection">
660 <a name="iterate_convert">Turning an iterator into a class pointer (and
664 <div class="doc_text">
666 <p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
667 instance when all you've got at hand is an iterator. Well, extracting
668 a reference or a pointer from an iterator is very straight-forward.
669 Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
670 is a <tt>BasicBlock::const_iterator</tt>:</p>
672 <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>
674 <p>However, the iterators you'll be working with in the LLVM framework are
675 special: they will automatically convert to a ptr-to-instance type whenever they
676 need to. Instead of dereferencing the iterator and then taking the address of
677 the result, you can simply assign the iterator to the proper pointer type and
678 you get the dereference and address-of operation as a result of the assignment
679 (behind the scenes, this is a result of overloading casting mechanisms). Thus
680 the last line of the last example,</p>
682 <pre>Instruction* pinst = &*i;</pre>
684 <p>is semantically equivalent to</p>
686 <pre>Instruction* pinst = i;</pre>
688 <p>It's also possible to turn a class pointer into the corresponding iterator,
689 and this is a constant time operation (very efficient). The following code
690 snippet illustrates use of the conversion constructors provided by LLVM
691 iterators. By using these, you can explicitly grab the iterator of something
692 without actually obtaining it via iteration over some structure:</p>
694 <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>
698 <!--_______________________________________________________________________-->
699 <div class="doc_subsubsection">
700 <a name="iterate_complex">Finding call sites: a slightly more complex
704 <div class="doc_text">
706 <p>Say that you're writing a FunctionPass and would like to count all the
707 locations in the entire module (that is, across every <tt>Function</tt>) where a
708 certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
709 learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
710 much more straight-forward manner, but this example will allow us to explore how
711 you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudocode, this
712 is what we want to do:</p>
714 <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>
716 <p>And the actual code is (remember, since we're writing a
717 <tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
718 override the <tt>runOnFunction</tt> method...):</p>
720 <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
721 href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a><<a
722 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>
726 <!--_______________________________________________________________________-->
727 <div class="doc_subsubsection">
728 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
731 <div class="doc_text">
733 <p>You may have noticed that the previous example was a bit oversimplified in
734 that it did not deal with call sites generated by 'invoke' instructions. In
735 this, and in other situations, you may find that you want to treat
736 <tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
737 most-specific common base class is <tt>Instruction</tt>, which includes lots of
738 less closely-related things. For these cases, LLVM provides a handy wrapper
740 href="http://llvm.cs.uiuc.edu/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
741 It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
742 methods that provide functionality common to <tt>CallInst</tt>s and
743 <tt>InvokeInst</tt>s.</p>
745 <p>This class has "value semantics": it should be passed by value, not by
746 reference and it should not be dynamically allocated or deallocated using
747 <tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
748 assignable and constructable, with costs equivalents to that of a bare pointer.
749 If you look at its definition, it has only a single pointer member.</p>
753 <!--_______________________________________________________________________-->
754 <div class="doc_subsubsection">
755 <a name="iterate_chains">Iterating over def-use & use-def chains</a>
758 <div class="doc_text">
760 <p>Frequently, we might have an instance of the <a
761 href="/doxygen/structllvm_1_1Value.html">Value Class</a> and we want to
762 determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
763 <tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
764 For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
765 particular function <tt>foo</tt>. Finding all of the instructions that
766 <i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
769 <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>
771 <p>Alternately, it's common to have an instance of the <a
772 href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
773 <tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
774 <tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
775 <tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
776 all of the values that a particular instruction uses (that is, the operands of
777 the particular <tt>Instruction</tt>):</p>
779 <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>
782 def-use chains ("finding all users of"): Value::use_begin/use_end
783 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
788 <!-- ======================================================================= -->
789 <div class="doc_subsection">
790 <a name="simplechanges">Making simple changes</a>
793 <div class="doc_text">
795 <p>There are some primitive transformation operations present in the LLVM
796 infrastructure that are worth knowing about. When performing
797 transformations, it's fairly common to manipulate the contents of basic
798 blocks. This section describes some of the common methods for doing so
799 and gives example code.</p>
803 <!--_______________________________________________________________________-->
804 <div class="doc_subsubsection">
805 <a name="schanges_creating">Creating and inserting new
806 <tt>Instruction</tt>s</a>
809 <div class="doc_text">
811 <p><i>Instantiating Instructions</i></p>
813 <p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
814 constructor for the kind of instruction to instantiate and provide the necessary
815 parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
816 (const-ptr-to) <tt>Type</tt>. Thus:</p>
818 <pre>AllocaInst* ai = new AllocaInst(Type::IntTy);</pre>
820 <p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
821 one integer in the current stack frame, at runtime. Each <tt>Instruction</tt>
822 subclass is likely to have varying default parameters which change the semantics
823 of the instruction, so refer to the <a
824 href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
825 Instruction</a> that you're interested in instantiating.</p>
827 <p><i>Naming values</i></p>
829 <p>It is very useful to name the values of instructions when you're able to, as
830 this facilitates the debugging of your transformations. If you end up looking
831 at generated LLVM machine code, you definitely want to have logical names
832 associated with the results of instructions! By supplying a value for the
833 <tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
834 associate a logical name with the result of the instruction's execution at
835 runtime. For example, say that I'm writing a transformation that dynamically
836 allocates space for an integer on the stack, and that integer is going to be
837 used as some kind of index by some other code. To accomplish this, I place an
838 <tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
839 <tt>Function</tt>, and I'm intending to use it within the same
840 <tt>Function</tt>. I might do:</p>
842 <pre>AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");</pre>
844 <p>where <tt>indexLoc</tt> is now the logical name of the instruction's
845 execution value, which is a pointer to an integer on the runtime stack.</p>
847 <p><i>Inserting instructions</i></p>
849 <p>There are essentially two ways to insert an <tt>Instruction</tt>
850 into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
853 <li>Insertion into an explicit instruction list
855 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
856 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
857 before <tt>*pi</tt>, we do the following: </p>
859 <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>
861 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
862 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
863 classes provide constructors which take a pointer to a
864 <tt>BasicBlock</tt> to be appended to. For example code that
867 <pre> BasicBlock *pb = ...;<br> Instruction *newInst = new Instruction(...);<br> pb->getInstList().push_back(newInst); // appends newInst to pb<br></pre>
871 <pre> BasicBlock *pb = ...;<br> Instruction *newInst = new Instruction(..., pb);<br></pre>
873 <p>which is much cleaner, especially if you are creating
874 long instruction streams.</p></li>
876 <li>Insertion into an implicit instruction list
878 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
879 are implicitly associated with an existing instruction list: the instruction
880 list of the enclosing basic block. Thus, we could have accomplished the same
881 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
884 <pre> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pi->getParent()->getInstList().insert(pi, newInst);<br></pre>
886 <p>In fact, this sequence of steps occurs so frequently that the
887 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
888 constructors which take (as a default parameter) a pointer to an
889 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
890 precede. That is, <tt>Instruction</tt> constructors are capable of
891 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
892 provided instruction, immediately before that instruction. Using an
893 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
894 parameter, the above code becomes:</p>
896 <pre>Instruction* pi = ...;<br>Instruction* newInst = new Instruction(..., pi);<br></pre>
898 <p>which is much cleaner, especially if you're creating a lot of
899 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
904 <!--_______________________________________________________________________-->
905 <div class="doc_subsubsection">
906 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
909 <div class="doc_text">
911 <p>Deleting an instruction from an existing sequence of instructions that form a
912 <a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
913 you must have a pointer to the instruction that you wish to delete. Second, you
914 need to obtain the pointer to that instruction's basic block. You use the
915 pointer to the basic block to get its list of instructions and then use the
916 erase function to remove your instruction. For example:</p>
918 <pre> <a href="#Instruction">Instruction</a> *I = .. ;<br> <a
919 href="#BasicBlock">BasicBlock</a> *BB = I->getParent();<br> BB->getInstList().erase(I);<br></pre>
923 <!--_______________________________________________________________________-->
924 <div class="doc_subsubsection">
925 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
929 <div class="doc_text">
931 <p><i>Replacing individual instructions</i></p>
933 <p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
934 permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
935 and <tt>ReplaceInstWithInst</tt>.</p>
937 <h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
940 <li><tt>ReplaceInstWithValue</tt>
942 <p>This function replaces all uses (within a basic block) of a given
943 instruction with a value, and then removes the original instruction. The
944 following example illustrates the replacement of the result of a particular
945 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
946 pointer to an integer.</p>
948 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithValue(instToReplace->getParent()->getInstList(), ii,<br> Constant::getNullValue(PointerType::get(Type::IntTy)));<br></pre></li>
950 <li><tt>ReplaceInstWithInst</tt>
952 <p>This function replaces a particular instruction with another
953 instruction. The following example illustrates the replacement of one
954 <tt>AllocaInst</tt> with another.</p>
956 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithInst(instToReplace->getParent()->getInstList(), ii,<br> new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));<br></pre></li>
959 <p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
961 <p>You can use <tt>Value::replaceAllUsesWith</tt> and
962 <tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
963 doxygen documentation for the <a href="/doxygen/structllvm_1_1Value.html">Value Class</a>
964 and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
967 <!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
968 include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
969 ReplaceInstWithValue, ReplaceInstWithInst -->
973 <!-- *********************************************************************** -->
974 <div class="doc_section">
975 <a name="advanced">Advanced Topics</a>
977 <!-- *********************************************************************** -->
979 <div class="doc_text">
981 This section describes some of the advanced or obscure API's that most clients
982 do not need to be aware of. These API's tend manage the inner workings of the
983 LLVM system, and only need to be accessed in unusual circumstances.
987 <!-- ======================================================================= -->
988 <div class="doc_subsection">
989 <a name="TypeResolve">LLVM Type Resolution</a>
992 <div class="doc_text">
995 The LLVM type system has a very simple goal: allow clients to compare types for
996 structural equality with a simple pointer comparison (aka a shallow compare).
997 This goal makes clients much simpler and faster, and is used throughout the LLVM
1002 Unfortunately achieving this goal is not a simple matter. In particular,
1003 recursive types and late resolution of opaque types makes the situation very
1004 difficult to handle. Fortunately, for the most part, our implementation makes
1005 most clients able to be completely unaware of the nasty internal details. The
1006 primary case where clients are exposed to the inner workings of it are when
1007 building a recursive type. In addition to this case, the LLVM bytecode reader,
1008 assembly parser, and linker also have to be aware of the inner workings of this
1013 For our purposes below, we need three concepts. First, an "Opaque Type" is
1014 exactly as defined in the <a href="LangRef.html#t_opaque">language
1015 reference</a>. Second an "Abstract Type" is any type which includes an
1016 opaque type as part of its type graph (for example "<tt>{ opaque, int }</tt>").
1017 Third, a concrete type is a type that is not an abstract type (e.g. "<tt>[ int,
1023 <!-- ______________________________________________________________________ -->
1024 <div class="doc_subsubsection">
1025 <a name="BuildRecType">Basic Recursive Type Construction</a>
1028 <div class="doc_text">
1031 Because the most common question is "how do I build a recursive type with LLVM",
1032 we answer it now and explain it as we go. Here we include enough to cause this
1033 to be emitted to an output .ll file:
1037 %mylist = type { %mylist*, int }
1041 To build this, use the following LLVM APIs:
1045 //<i> Create the initial outer struct.</i>
1046 <a href="#PATypeHolder">PATypeHolder</a> StructTy = OpaqueType::get();
1047 std::vector<const Type*> Elts;
1048 Elts.push_back(PointerType::get(StructTy));
1049 Elts.push_back(Type::IntTy);
1050 StructType *NewSTy = StructType::get(Elts);
1052 //<i> At this point, NewSTy = "{ opaque*, int }". Tell VMCore that</i>
1053 //<i> the struct and the opaque type are actually the same.</i>
1054 cast<OpaqueType>(StructTy.get())-><a href="#refineAbstractTypeTo">refineAbstractTypeTo</a>(NewSTy);
1056 // <i>NewSTy is potentially invalidated, but StructTy (a <a href="#PATypeHolder">PATypeHolder</a>) is</i>
1057 // <i>kept up-to-date.</i>
1058 NewSTy = cast<StructType>(StructTy.get());
1060 // <i>Add a name for the type to the module symbol table (optional).</i>
1061 MyModule->addTypeName("mylist", NewSTy);
1065 This code shows the basic approach used to build recursive types: build a
1066 non-recursive type using 'opaque', then use type unification to close the cycle.
1067 The type unification step is performed by the <tt><a
1068 ref="#refineAbstractTypeTo">refineAbstractTypeTo</a></tt> method, which is
1069 described next. After that, we describe the <a
1070 href="#PATypeHolder">PATypeHolder class</a>.
1075 <!-- ______________________________________________________________________ -->
1076 <div class="doc_subsubsection">
1077 <a name="refineAbstractTypeTo">The <tt>refineAbstractTypeTo</tt> method</a>
1080 <div class="doc_text">
1082 The <tt>refineAbstractTypeTo</tt> method starts the type unification process.
1083 While this method is actually a member of the DerivedType class, it is most
1084 often used on OpaqueType instances. Type unification is actually a recursive
1085 process. After unification, types can become structurally isomorphic to
1086 existing types, and all duplicates are deleted (to preserve pointer equality).
1090 In the example above, the OpaqueType object is definitely deleted.
1091 Additionally, if there is an "{ \2*, int}" type already created in the system,
1092 the pointer and struct type created are <b>also</b> deleted. Obviously whenever
1093 a type is deleted, any "Type*" pointers in the program are invalidated. As
1094 such, it is safest to avoid having <i>any</i> "Type*" pointers to abstract types
1095 live across a call to <tt>refineAbstractTypeTo</tt> (note that non-abstract
1096 types can never move or be deleted). To deal with this, the <a
1097 href="#PATypeHolder">PATypeHolder</a> class is used to maintain a stable
1098 reference to a possibly refined type, and the <a
1099 href="#AbstractTypeUser">AbstractTypeUser</a> class is used to update more
1100 complex datastructures.
1105 <!-- ______________________________________________________________________ -->
1106 <div class="doc_subsubsection">
1107 <a name="PATypeHolder">The PATypeHolder Class</a>
1110 <div class="doc_text">
1112 PATypeHolder is a form of a "smart pointer" for Type objects. When VMCore
1113 happily goes about nuking types that become isomorphic to existing types, it
1114 automatically updates all PATypeHolder objects to point to the new type. In the
1115 example above, this allows the code to maintain a pointer to the resultant
1116 resolved recursive type, even though the Type*'s are potentially invalidated.
1120 PATypeHolder is an extremely light-weight object that uses a lazy union-find
1121 implementation to update pointers. For example the pointer from a Value to its
1122 Type is maintained by PATypeHolder objects.
1127 <!-- ______________________________________________________________________ -->
1128 <div class="doc_subsubsection">
1129 <a name="AbstractTypeUser">The AbstractTypeUser Class</a>
1132 <div class="doc_text">
1135 Some data structures need more to perform more complex updates when types get
1136 resolved. The <a href="#SymbolTable">SymbolTable</a> class, for example, needs
1137 move and potentially merge type planes in its representation when a pointer
1141 To support this, a class can derive from the AbstractTypeUser class. This class
1142 allows it to get callbacks when certain types are resolved. To register to get
1143 callbacks for a particular type, the DerivedType::{add/remove}AbstractTypeUser
1144 methods can be called on a type. Note that these methods only work for <i>
1145 abstract</i> types. Concrete types (those that do not include an opaque objects
1146 somewhere) can never be refined.
1151 <!-- ======================================================================= -->
1152 <div class="doc_subsection">
1153 <a name="SymbolTable">The <tt>SymbolTable</tt> class</a>
1156 <div class="doc_text">
1157 <p>This class provides a symbol table that the <a
1158 href="#Function"><tt>Function</tt></a> and <a href="#Module">
1159 <tt>Module</tt></a> classes use for naming definitions. The symbol table can
1160 provide a name for any <a href="#Value"><tt>Value</tt></a> or <a
1161 href="#Type"><tt>Type</tt></a>. <tt>SymbolTable</tt> is an abstract data
1162 type. It hides the data it contains and provides access to it through a
1163 controlled interface.</p>
1165 <p>Note that the symbol table class is should not be directly accessed by most
1166 clients. It should only be used when iteration over the symbol table names
1167 themselves are required, which is very special purpose. Note that not all LLVM
1168 <a href="#Value">Value</a>s have names, and those without names (i.e. they have
1169 an empty name) do not exist in the symbol table.
1172 <p>To use the <tt>SymbolTable</tt> well, you need to understand the
1173 structure of the information it holds. The class contains two
1174 <tt>std::map</tt> objects. The first, <tt>pmap</tt>, is a map of
1175 <tt>Type*</tt> to maps of name (<tt>std::string</tt>) to <tt>Value*</tt>.
1176 The second, <tt>tmap</tt>, is a map of names to <tt>Type*</tt>. Thus, Values
1177 are stored in two-dimensions and accessed by <tt>Type</tt> and name. Types,
1178 however, are stored in a single dimension and accessed only by name.</p>
1180 <p>The interface of this class provides three basic types of operations:
1182 <li><em>Accessors</em>. Accessors provide read-only access to information
1183 such as finding a value for a name with the
1184 <a href="#SymbolTable_lookup">lookup</a> method.</li>
1185 <li><em>Mutators</em>. Mutators allow the user to add information to the
1186 <tt>SymbolTable</tt> with methods like
1187 <a href="#SymbolTable_insert"><tt>insert</tt></a>.</li>
1188 <li><em>Iterators</em>. Iterators allow the user to traverse the content
1189 of the symbol table in well defined ways, such as the method
1190 <a href="#SymbolTable_type_begin"><tt>type_begin</tt></a>.</li>
1195 <dt><tt>Value* lookup(const Type* Ty, const std::string& name) const</tt>:
1197 <dd>The <tt>lookup</tt> method searches the type plane given by the
1198 <tt>Ty</tt> parameter for a <tt>Value</tt> with the provided <tt>name</tt>.
1199 If a suitable <tt>Value</tt> is not found, null is returned.</dd>
1201 <dt><tt>Type* lookupType( const std::string& name) const</tt>:</dt>
1202 <dd>The <tt>lookupType</tt> method searches through the types for a
1203 <tt>Type</tt> with the provided <tt>name</tt>. If a suitable <tt>Type</tt>
1204 is not found, null is returned.</dd>
1206 <dt><tt>bool hasTypes() const</tt>:</dt>
1207 <dd>This function returns true if an entry has been made into the type
1210 <dt><tt>bool isEmpty() const</tt>:</dt>
1211 <dd>This function returns true if both the value and types maps are
1217 <dt><tt>void insert(Value *Val)</tt>:</dt>
1218 <dd>This method adds the provided value to the symbol table. The Value must
1219 have both a name and a type which are extracted and used to place the value
1220 in the correct type plane under the value's name.</dd>
1222 <dt><tt>void insert(const std::string& Name, Value *Val)</tt>:</dt>
1223 <dd> Inserts a constant or type into the symbol table with the specified
1224 name. There can be a many to one mapping between names and constants
1227 <dt><tt>void insert(const std::string& Name, Type *Typ)</tt>:</dt>
1228 <dd> Inserts a type into the symbol table with the specified name. There
1229 can be a many-to-one mapping between names and types. This method
1230 allows a type with an existing entry in the symbol table to get
1233 <dt><tt>void remove(Value* Val)</tt>:</dt>
1234 <dd> This method removes a named value from the symbol table. The
1235 type and name of the Value are extracted from \p N and used to
1236 lookup the Value in the correct type plane. If the Value is
1237 not in the symbol table, this method silently ignores the
1240 <dt><tt>void remove(Type* Typ)</tt>:</dt>
1241 <dd> This method removes a named type from the symbol table. The
1242 name of the type is extracted from \P T and used to look up
1243 the Type in the type map. If the Type is not in the symbol
1244 table, this method silently ignores the request.</dd>
1246 <dt><tt>Value* remove(const std::string& Name, Value *Val)</tt>:</dt>
1247 <dd> Remove a constant or type with the specified name from the
1250 <dt><tt>Type* remove(const std::string& Name, Type* T)</tt>:</dt>
1251 <dd> Remove a type with the specified name from the symbol table.
1252 Returns the removed Type.</dd>
1254 <dt><tt>Value *value_remove(const value_iterator& It)</tt>:</dt>
1255 <dd> Removes a specific value from the symbol table.
1256 Returns the removed value.</dd>
1258 <dt><tt>bool strip()</tt>:</dt>
1259 <dd> This method will strip the symbol table of its names leaving
1260 the type and values. </dd>
1262 <dt><tt>void clear()</tt>:</dt>
1263 <dd>Empty the symbol table completely.</dd>
1267 <p>The following functions describe three types of iterators you can obtain
1268 the beginning or end of the sequence for both const and non-const. It is
1269 important to keep track of the different kinds of iterators. There are
1270 three idioms worth pointing out:</p>
1272 <tr><th>Units</th><th>Iterator</th><th>Idiom</th></tr>
1274 <td align="left">Planes Of name/Value maps</td><td>PI</td>
1275 <td align="left"><pre><tt>
1276 for (SymbolTable::plane_const_iterator PI = ST.plane_begin(),
1277 PE = ST.plane_end(); PI != PE; ++PI ) {
1278 PI->first // This is the Type* of the plane
1279 PI->second // This is the SymbolTable::ValueMap of name/Value pairs
1283 <td align="left">All name/Type Pairs</td><td>TI</td>
1284 <td align="left"><pre><tt>
1285 for (SymbolTable::type_const_iterator TI = ST.type_begin(),
1286 TE = ST.type_end(); TI != TE; ++TI )
1287 TI->first // This is the name of the type
1288 TI->second // This is the Type* value associated with the name
1292 <td align="left">name/Value pairs in a plane</td><td>VI</td>
1293 <td align="left"><pre><tt>
1294 for (SymbolTable::value_const_iterator VI = ST.value_begin(SomeType),
1295 VE = ST.value_end(SomeType); VI != VE; ++VI )
1296 VI->first // This is the name of the Value
1297 VI->second // This is the Value* value associated with the name
1302 <p>Using the recommended iterator names and idioms will help you avoid
1303 making mistakes. Of particular note, make sure that whenever you use
1304 value_begin(SomeType) that you always compare the resulting iterator
1305 with value_end(SomeType) not value_end(SomeOtherType) or else you
1306 will loop infinitely.</p>
1310 <dt><tt>plane_iterator plane_begin()</tt>:</dt>
1311 <dd>Get an iterator that starts at the beginning of the type planes.
1312 The iterator will iterate over the Type/ValueMap pairs in the
1315 <dt><tt>plane_const_iterator plane_begin() const</tt>:</dt>
1316 <dd>Get a const_iterator that starts at the beginning of the type
1317 planes. The iterator will iterate over the Type/ValueMap pairs
1318 in the type planes. </dd>
1320 <dt><tt>plane_iterator plane_end()</tt>:</dt>
1321 <dd>Get an iterator at the end of the type planes. This serves as
1322 the marker for end of iteration over the type planes.</dd>
1324 <dt><tt>plane_const_iterator plane_end() const</tt>:</dt>
1325 <dd>Get a const_iterator at the end of the type planes. This serves as
1326 the marker for end of iteration over the type planes.</dd>
1328 <dt><tt>value_iterator value_begin(const Type *Typ)</tt>:</dt>
1329 <dd>Get an iterator that starts at the beginning of a type plane.
1330 The iterator will iterate over the name/value pairs in the type plane.
1331 Note: The type plane must already exist before using this.</dd>
1333 <dt><tt>value_const_iterator value_begin(const Type *Typ) const</tt>:</dt>
1334 <dd>Get a const_iterator that starts at the beginning of a type plane.
1335 The iterator will iterate over the name/value pairs in the type plane.
1336 Note: The type plane must already exist before using this.</dd>
1338 <dt><tt>value_iterator value_end(const Type *Typ)</tt>:</dt>
1339 <dd>Get an iterator to the end of a type plane. This serves as the marker
1340 for end of iteration of the type plane.
1341 Note: The type plane must already exist before using this.</dd>
1343 <dt><tt>value_const_iterator value_end(const Type *Typ) const</tt>:</dt>
1344 <dd>Get a const_iterator to the end of a type plane. This serves as the
1345 marker for end of iteration of the type plane.
1346 Note: the type plane must already exist before using this.</dd>
1348 <dt><tt>type_iterator type_begin()</tt>:</dt>
1349 <dd>Get an iterator to the start of the name/Type map.</dd>
1351 <dt><tt>type_const_iterator type_begin() cons</tt>:</dt>
1352 <dd> Get a const_iterator to the start of the name/Type map.</dd>
1354 <dt><tt>type_iterator type_end()</tt>:</dt>
1355 <dd>Get an iterator to the end of the name/Type map. This serves as the
1356 marker for end of iteration of the types.</dd>
1358 <dt><tt>type_const_iterator type_end() const</tt>:</dt>
1359 <dd>Get a const-iterator to the end of the name/Type map. This serves
1360 as the marker for end of iteration of the types.</dd>
1362 <dt><tt>plane_const_iterator find(const Type* Typ ) const</tt>:</dt>
1363 <dd>This method returns a plane_const_iterator for iteration over
1364 the type planes starting at a specific plane, given by \p Ty.</dd>
1366 <dt><tt>plane_iterator find( const Type* Typ </tt>:</dt>
1367 <dd>This method returns a plane_iterator for iteration over the
1368 type planes starting at a specific plane, given by \p Ty.</dd>
1375 <!-- *********************************************************************** -->
1376 <div class="doc_section">
1377 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
1379 <!-- *********************************************************************** -->
1381 <div class="doc_text">
1383 <p>The Core LLVM classes are the primary means of representing the program
1384 being inspected or transformed. The core LLVM classes are defined in
1385 header files in the <tt>include/llvm/</tt> directory, and implemented in
1386 the <tt>lib/VMCore</tt> directory.</p>
1390 <!-- ======================================================================= -->
1391 <div class="doc_subsection">
1392 <a name="Value">The <tt>Value</tt> class</a>
1397 <p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
1399 doxygen info: <a href="/doxygen/structllvm_1_1Value.html">Value Class</a></p>
1401 <p>The <tt>Value</tt> class is the most important class in the LLVM Source
1402 base. It represents a typed value that may be used (among other things) as an
1403 operand to an instruction. There are many different types of <tt>Value</tt>s,
1404 such as <a href="#Constant"><tt>Constant</tt></a>s,<a
1405 href="#Argument"><tt>Argument</tt></a>s. Even <a
1406 href="#Instruction"><tt>Instruction</tt></a>s and <a
1407 href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
1409 <p>A particular <tt>Value</tt> may be used many times in the LLVM representation
1410 for a program. For example, an incoming argument to a function (represented
1411 with an instance of the <a href="#Argument">Argument</a> class) is "used" by
1412 every instruction in the function that references the argument. To keep track
1413 of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
1414 href="#User"><tt>User</tt></a>s that is using it (the <a
1415 href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
1416 graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
1417 def-use information in the program, and is accessible through the <tt>use_</tt>*
1418 methods, shown below.</p>
1420 <p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
1421 and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
1422 method. In addition, all LLVM values can be named. The "name" of the
1423 <tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
1425 <pre> %<b>foo</b> = add int 1, 2<br></pre>
1427 <p><a name="#nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
1428 that the name of any value may be missing (an empty string), so names should
1429 <b>ONLY</b> be used for debugging (making the source code easier to read,
1430 debugging printouts), they should not be used to keep track of values or map
1431 between them. For this purpose, use a <tt>std::map</tt> of pointers to the
1432 <tt>Value</tt> itself instead.</p>
1434 <p>One important aspect of LLVM is that there is no distinction between an SSA
1435 variable and the operation that produces it. Because of this, any reference to
1436 the value produced by an instruction (or the value available as an incoming
1437 argument, for example) is represented as a direct pointer to the instance of
1439 represents this value. Although this may take some getting used to, it
1440 simplifies the representation and makes it easier to manipulate.</p>
1444 <!-- _______________________________________________________________________ -->
1445 <div class="doc_subsubsection">
1446 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
1449 <div class="doc_text">
1452 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
1454 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
1456 <tt>unsigned use_size()</tt> - Returns the number of users of the
1458 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
1459 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
1461 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
1463 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
1464 element in the list.
1465 <p> These methods are the interface to access the def-use
1466 information in LLVM. As with all other iterators in LLVM, the naming
1467 conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
1469 <li><tt><a href="#Type">Type</a> *getType() const</tt>
1470 <p>This method returns the Type of the Value.</p>
1472 <li><tt>bool hasName() const</tt><br>
1473 <tt>std::string getName() const</tt><br>
1474 <tt>void setName(const std::string &Name)</tt>
1475 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
1476 be aware of the <a href="#nameWarning">precaution above</a>.</p>
1478 <li><tt>void replaceAllUsesWith(Value *V)</tt>
1480 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
1481 href="#User"><tt>User</tt>s</a> of the current value to refer to
1482 "<tt>V</tt>" instead. For example, if you detect that an instruction always
1483 produces a constant value (for example through constant folding), you can
1484 replace all uses of the instruction with the constant like this:</p>
1486 <pre> Inst->replaceAllUsesWith(ConstVal);<br></pre>
1491 <!-- ======================================================================= -->
1492 <div class="doc_subsection">
1493 <a name="User">The <tt>User</tt> class</a>
1496 <div class="doc_text">
1499 <tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
1500 doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
1501 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1503 <p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
1504 refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
1505 that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
1506 referring to. The <tt>User</tt> class itself is a subclass of
1509 <p>The operands of a <tt>User</tt> point directly to the LLVM <a
1510 href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
1511 Single Assignment (SSA) form, there can only be one definition referred to,
1512 allowing this direct connection. This connection provides the use-def
1513 information in LLVM.</p>
1517 <!-- _______________________________________________________________________ -->
1518 <div class="doc_subsubsection">
1519 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
1522 <div class="doc_text">
1524 <p>The <tt>User</tt> class exposes the operand list in two ways: through
1525 an index access interface and through an iterator based interface.</p>
1528 <li><tt>Value *getOperand(unsigned i)</tt><br>
1529 <tt>unsigned getNumOperands()</tt>
1530 <p> These two methods expose the operands of the <tt>User</tt> in a
1531 convenient form for direct access.</p></li>
1533 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
1535 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
1536 the operand list.<br>
1537 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
1539 <p> Together, these methods make up the iterator based interface to
1540 the operands of a <tt>User</tt>.</p></li>
1545 <!-- ======================================================================= -->
1546 <div class="doc_subsection">
1547 <a name="Instruction">The <tt>Instruction</tt> class</a>
1550 <div class="doc_text">
1552 <p><tt>#include "</tt><tt><a
1553 href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
1554 doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
1555 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1556 href="#Value"><tt>Value</tt></a></p>
1558 <p>The <tt>Instruction</tt> class is the common base class for all LLVM
1559 instructions. It provides only a few methods, but is a very commonly used
1560 class. The primary data tracked by the <tt>Instruction</tt> class itself is the
1561 opcode (instruction type) and the parent <a
1562 href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
1563 into. To represent a specific type of instruction, one of many subclasses of
1564 <tt>Instruction</tt> are used.</p>
1566 <p> Because the <tt>Instruction</tt> class subclasses the <a
1567 href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
1568 way as for other <a href="#User"><tt>User</tt></a>s (with the
1569 <tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
1570 <tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
1571 the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
1572 file contains some meta-data about the various different types of instructions
1573 in LLVM. It describes the enum values that are used as opcodes (for example
1574 <tt>Instruction::Add</tt> and <tt>Instruction::SetLE</tt>), as well as the
1575 concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
1576 example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
1577 href="#SetCondInst">SetCondInst</a></tt>). Unfortunately, the use of macros in
1578 this file confuses doxygen, so these enum values don't show up correctly in the
1579 <a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
1583 <!-- _______________________________________________________________________ -->
1584 <div class="doc_subsubsection">
1585 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
1589 <div class="doc_text">
1592 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
1593 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
1594 this <tt>Instruction</tt> is embedded into.</p></li>
1595 <li><tt>bool mayWriteToMemory()</tt>
1596 <p>Returns true if the instruction writes to memory, i.e. it is a
1597 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
1598 <li><tt>unsigned getOpcode()</tt>
1599 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
1600 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
1601 <p>Returns another instance of the specified instruction, identical
1602 in all ways to the original except that the instruction has no parent
1603 (ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
1604 and it has no name</p></li>
1609 <!-- ======================================================================= -->
1610 <div class="doc_subsection">
1611 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
1614 <div class="doc_text">
1617 href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
1618 doxygen info: <a href="/doxygen/structllvm_1_1BasicBlock.html">BasicBlock
1620 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1622 <p>This class represents a single entry multiple exit section of the code,
1623 commonly known as a basic block by the compiler community. The
1624 <tt>BasicBlock</tt> class maintains a list of <a
1625 href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
1626 Matching the language definition, the last element of this list of instructions
1627 is always a terminator instruction (a subclass of the <a
1628 href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
1630 <p>In addition to tracking the list of instructions that make up the block, the
1631 <tt>BasicBlock</tt> class also keeps track of the <a
1632 href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
1634 <p>Note that <tt>BasicBlock</tt>s themselves are <a
1635 href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
1636 like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
1641 <!-- _______________________________________________________________________ -->
1642 <div class="doc_subsubsection">
1643 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
1647 <div class="doc_text">
1651 <li><tt>BasicBlock(const std::string &Name = "", </tt><tt><a
1652 href="#Function">Function</a> *Parent = 0)</tt>
1654 <p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
1655 insertion into a function. The constructor optionally takes a name for the new
1656 block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
1657 the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
1658 automatically inserted at the end of the specified <a
1659 href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
1660 manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
1662 <li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
1663 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
1664 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1665 <tt>size()</tt>, <tt>empty()</tt>
1666 STL-style functions for accessing the instruction list.
1668 <p>These methods and typedefs are forwarding functions that have the same
1669 semantics as the standard library methods of the same names. These methods
1670 expose the underlying instruction list of a basic block in a way that is easy to
1671 manipulate. To get the full complement of container operations (including
1672 operations to update the list), you must use the <tt>getInstList()</tt>
1675 <li><tt>BasicBlock::InstListType &getInstList()</tt>
1677 <p>This method is used to get access to the underlying container that actually
1678 holds the Instructions. This method must be used when there isn't a forwarding
1679 function in the <tt>BasicBlock</tt> class for the operation that you would like
1680 to perform. Because there are no forwarding functions for "updating"
1681 operations, you need to use this if you want to update the contents of a
1682 <tt>BasicBlock</tt>.</p></li>
1684 <li><tt><a href="#Function">Function</a> *getParent()</tt>
1686 <p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
1687 embedded into, or a null pointer if it is homeless.</p></li>
1689 <li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
1691 <p> Returns a pointer to the terminator instruction that appears at the end of
1692 the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
1693 instruction in the block is not a terminator, then a null pointer is
1700 <!-- ======================================================================= -->
1701 <div class="doc_subsection">
1702 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
1705 <div class="doc_text">
1708 href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
1709 doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
1711 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1712 href="#Value"><tt>Value</tt></a></p>
1714 <p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
1715 href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
1716 visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
1717 Because they are visible at global scope, they are also subject to linking with
1718 other globals defined in different translation units. To control the linking
1719 process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
1720 <tt>GlobalValue</tt>s know whether they have internal or external linkage, as
1721 defined by the <tt>LinkageTypes</tt> enumeration.</p>
1723 <p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
1724 <tt>static</tt> in C), it is not visible to code outside the current translation
1725 unit, and does not participate in linking. If it has external linkage, it is
1726 visible to external code, and does participate in linking. In addition to
1727 linkage information, <tt>GlobalValue</tt>s keep track of which <a
1728 href="#Module"><tt>Module</tt></a> they are currently part of.</p>
1730 <p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
1731 by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
1732 global is always a pointer to its contents. It is important to remember this
1733 when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
1734 be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
1735 subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
1736 int]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
1737 the address of the first element of this array and the value of the
1738 <tt>GlobalVariable</tt> are the same, they have different types. The
1739 <tt>GlobalVariable</tt>'s type is <tt>[24 x int]</tt>. The first element's type
1740 is <tt>int.</tt> Because of this, accessing a global value requires you to
1741 dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
1742 can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
1743 Language Reference Manual</a>.</p>
1747 <!-- _______________________________________________________________________ -->
1748 <div class="doc_subsubsection">
1749 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
1753 <div class="doc_text">
1756 <li><tt>bool hasInternalLinkage() const</tt><br>
1757 <tt>bool hasExternalLinkage() const</tt><br>
1758 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
1759 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
1762 <li><tt><a href="#Module">Module</a> *getParent()</tt>
1763 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
1764 GlobalValue is currently embedded into.</p></li>
1769 <!-- ======================================================================= -->
1770 <div class="doc_subsection">
1771 <a name="Function">The <tt>Function</tt> class</a>
1774 <div class="doc_text">
1777 href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
1778 info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
1779 Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1780 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1782 <p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
1783 actually one of the more complex classes in the LLVM heirarchy because it must
1784 keep track of a large amount of data. The <tt>Function</tt> class keeps track
1785 of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal <a
1786 href="#Argument"><tt>Argument</tt></a>s, and a <a
1787 href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
1789 <p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
1790 commonly used part of <tt>Function</tt> objects. The list imposes an implicit
1791 ordering of the blocks in the function, which indicate how the code will be
1792 layed out by the backend. Additionally, the first <a
1793 href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
1794 <tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
1795 block. There are no implicit exit nodes, and in fact there may be multiple exit
1796 nodes from a single <tt>Function</tt>. If the <a
1797 href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
1798 the <tt>Function</tt> is actually a function declaration: the actual body of the
1799 function hasn't been linked in yet.</p>
1801 <p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
1802 <tt>Function</tt> class also keeps track of the list of formal <a
1803 href="#Argument"><tt>Argument</tt></a>s that the function receives. This
1804 container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
1805 nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
1806 the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
1808 <p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
1809 LLVM feature that is only used when you have to look up a value by name. Aside
1810 from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
1811 internally to make sure that there are not conflicts between the names of <a
1812 href="#Instruction"><tt>Instruction</tt></a>s, <a
1813 href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
1814 href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
1816 <p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
1817 and therefore also a <a href="#Constant">Constant</a>. The value of the function
1818 is its address (after linking) which is guaranteed to be constant.</p>
1821 <!-- _______________________________________________________________________ -->
1822 <div class="doc_subsubsection">
1823 <a name="m_Function">Important Public Members of the <tt>Function</tt>
1827 <div class="doc_text">
1830 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
1831 *Ty, LinkageTypes Linkage, const std::string &N = "", Module* Parent = 0)</tt>
1833 <p>Constructor used when you need to create new <tt>Function</tt>s to add
1834 the the program. The constructor must specify the type of the function to
1835 create and what type of linkage the function should have. The <a
1836 href="#FunctionType"><tt>FunctionType</tt></a> argument
1837 specifies the formal arguments and return value for the function. The same
1838 <a href="#FunctionTypel"><tt>FunctionType</tt></a> value can be used to
1839 create multiple functions. The <tt>Parent</tt> argument specifies the Module
1840 in which the function is defined. If this argument is provided, the function
1841 will automatically be inserted into that module's list of
1844 <li><tt>bool isExternal()</tt>
1846 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
1847 function is "external", it does not have a body, and thus must be resolved
1848 by linking with a function defined in a different translation unit.</p></li>
1850 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
1851 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
1853 <tt>begin()</tt>, <tt>end()</tt>
1854 <tt>size()</tt>, <tt>empty()</tt>
1856 <p>These are forwarding methods that make it easy to access the contents of
1857 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
1860 <li><tt>Function::BasicBlockListType &getBasicBlockList()</tt>
1862 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
1863 is necessary to use when you need to update the list or perform a complex
1864 action that doesn't have a forwarding method.</p></li>
1866 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
1868 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
1870 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
1871 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
1873 <p>These are forwarding methods that make it easy to access the contents of
1874 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
1877 <li><tt>Function::ArgumentListType &getArgumentList()</tt>
1879 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
1880 necessary to use when you need to update the list or perform a complex
1881 action that doesn't have a forwarding method.</p></li>
1883 <li><tt><a href="#BasicBlock">BasicBlock</a> &getEntryBlock()</tt>
1885 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
1886 function. Because the entry block for the function is always the first
1887 block, this returns the first block of the <tt>Function</tt>.</p></li>
1889 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
1890 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
1892 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
1893 <tt>Function</tt> and returns the return type of the function, or the <a
1894 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
1897 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1899 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1900 for this <tt>Function</tt>.</p></li>
1905 <!-- ======================================================================= -->
1906 <div class="doc_subsection">
1907 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
1910 <div class="doc_text">
1913 href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
1915 doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
1916 Class</a><br> Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1917 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1919 <p>Global variables are represented with the (suprise suprise)
1920 <tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
1921 subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
1922 always referenced by their address (global values must live in memory, so their
1923 "name" refers to their address). See <a
1924 href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global variables
1925 may have an initial value (which must be a <a
1926 href="#Constant"><tt>Constant</tt></a>), and if they have an initializer, they
1927 may be marked as "constant" themselves (indicating that their contents never
1928 change at runtime).</p>
1932 <!-- _______________________________________________________________________ -->
1933 <div class="doc_subsubsection">
1934 <a name="m_GlobalVariable">Important Public Members of the
1935 <tt>GlobalVariable</tt> class</a>
1938 <div class="doc_text">
1941 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
1942 isConstant, LinkageTypes& Linkage, <a href="#Constant">Constant</a>
1943 *Initializer = 0, const std::string &Name = "", Module* Parent = 0)</tt>
1945 <p>Create a new global variable of the specified type. If
1946 <tt>isConstant</tt> is true then the global variable will be marked as
1947 unchanging for the program. The Linkage parameter specifies the type of
1948 linkage (internal, external, weak, linkonce, appending) for the variable. If
1949 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage, then
1950 the resultant global variable will have internal linkage. AppendingLinkage
1951 concatenates together all instances (in different translation units) of the
1952 variable into a single variable but is only applicable to arrays. See
1953 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
1954 further details on linkage types. Optionally an initializer, a name, and the
1955 module to put the variable into may be specified for the global variable as
1958 <li><tt>bool isConstant() const</tt>
1960 <p>Returns true if this is a global variable that is known not to
1961 be modified at runtime.</p></li>
1963 <li><tt>bool hasInitializer()</tt>
1965 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
1967 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
1969 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
1970 to call this method if there is no initializer.</p></li>
1975 <!-- ======================================================================= -->
1976 <div class="doc_subsection">
1977 <a name="Module">The <tt>Module</tt> class</a>
1980 <div class="doc_text">
1983 href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
1984 <a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
1986 <p>The <tt>Module</tt> class represents the top level structure present in LLVM
1987 programs. An LLVM module is effectively either a translation unit of the
1988 original program or a combination of several translation units merged by the
1989 linker. The <tt>Module</tt> class keeps track of a list of <a
1990 href="#Function"><tt>Function</tt></a>s, a list of <a
1991 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
1992 href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
1993 helpful member functions that try to make common operations easy.</p>
1997 <!-- _______________________________________________________________________ -->
1998 <div class="doc_subsubsection">
1999 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
2002 <div class="doc_text">
2005 <li><tt>Module::Module(std::string name = "")</tt></li>
2008 <p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
2009 provide a name for it (probably based on the name of the translation unit).</p>
2012 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
2013 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
2015 <tt>begin()</tt>, <tt>end()</tt>
2016 <tt>size()</tt>, <tt>empty()</tt>
2018 <p>These are forwarding methods that make it easy to access the contents of
2019 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
2022 <li><tt>Module::FunctionListType &getFunctionList()</tt>
2024 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
2025 necessary to use when you need to update the list or perform a complex
2026 action that doesn't have a forwarding method.</p>
2028 <p><!-- Global Variable --></p></li>
2034 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
2036 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
2038 <tt>global_begin()</tt>, <tt>global_end()</tt>
2039 <tt>global_size()</tt>, <tt>global_empty()</tt>
2041 <p> These are forwarding methods that make it easy to access the contents of
2042 a <tt>Module</tt> object's <a
2043 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
2045 <li><tt>Module::GlobalListType &getGlobalList()</tt>
2047 <p>Returns the list of <a
2048 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
2049 use when you need to update the list or perform a complex action that
2050 doesn't have a forwarding method.</p>
2052 <p><!-- Symbol table stuff --> </p></li>
2058 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
2060 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2061 for this <tt>Module</tt>.</p>
2063 <p><!-- Convenience methods --></p></li>
2069 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
2070 &Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
2072 <p>Look up the specified function in the <tt>Module</tt> <a
2073 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
2074 <tt>null</tt>.</p></li>
2076 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
2077 std::string &Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
2079 <p>Look up the specified function in the <tt>Module</tt> <a
2080 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
2081 external declaration for the function and return it.</p></li>
2083 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
2085 <p>If there is at least one entry in the <a
2086 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
2087 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
2090 <li><tt>bool addTypeName(const std::string &Name, const <a
2091 href="#Type">Type</a> *Ty)</tt>
2093 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
2094 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
2095 name, true is returned and the <a
2096 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
2101 <!-- ======================================================================= -->
2102 <div class="doc_subsection">
2103 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
2106 <div class="doc_text">
2108 <p>Constant represents a base class for different types of constants. It
2109 is subclassed by ConstantBool, ConstantInt, ConstantSInt, ConstantUInt,
2110 ConstantArray etc for representing the various types of Constants.</p>
2114 <!-- _______________________________________________________________________ -->
2115 <div class="doc_subsubsection">
2116 <a name="m_Constant">Important Public Methods</a>
2118 <div class="doc_text">
2121 <!-- _______________________________________________________________________ -->
2122 <div class="doc_subsubsection">Important Subclasses of Constant </div>
2123 <div class="doc_text">
2125 <li>ConstantSInt : This subclass of Constant represents a signed integer
2128 <li><tt>int64_t getValue() const</tt>: Returns the underlying value of
2129 this constant. </li>
2132 <li>ConstantUInt : This class represents an unsigned integer.
2134 <li><tt>uint64_t getValue() const</tt>: Returns the underlying value of
2135 this constant. </li>
2138 <li>ConstantFP : This class represents a floating point constant.
2140 <li><tt>double getValue() const</tt>: Returns the underlying value of
2141 this constant. </li>
2144 <li>ConstantBool : This represents a boolean constant.
2146 <li><tt>bool getValue() const</tt>: Returns the underlying value of this
2150 <li>ConstantArray : This represents a constant array.
2152 <li><tt>const std::vector<Use> &getValues() const</tt>: Returns
2153 a vector of component constants that makeup this array. </li>
2156 <li>ConstantStruct : This represents a constant struct.
2158 <li><tt>const std::vector<Use> &getValues() const</tt>: Returns
2159 a vector of component constants that makeup this array. </li>
2162 <li>GlobalValue : This represents either a global variable or a function. In
2163 either case, the value is a constant fixed address (after linking).
2168 <!-- ======================================================================= -->
2169 <div class="doc_subsection">
2170 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
2173 <div class="doc_text">
2175 <p>Type as noted earlier is also a subclass of a Value class. Any primitive
2176 type (like int, short etc) in LLVM is an instance of Type Class. All other
2177 types are instances of subclasses of type like FunctionType, ArrayType
2178 etc. DerivedType is the interface for all such dervied types including
2179 FunctionType, ArrayType, PointerType, StructType. Types can have names. They can
2180 be recursive (StructType). There exists exactly one instance of any type
2181 structure at a time. This allows using pointer equality of Type *s for comparing
2186 <!-- _______________________________________________________________________ -->
2187 <div class="doc_subsubsection">
2188 <a name="m_Value">Important Public Methods</a>
2191 <div class="doc_text">
2195 <li><tt>bool isSigned() const</tt>: Returns whether an integral numeric type
2196 is signed. This is true for SByteTy, ShortTy, IntTy, LongTy. Note that this is
2197 not true for Float and Double. </li>
2199 <li><tt>bool isUnsigned() const</tt>: Returns whether a numeric type is
2200 unsigned. This is not quite the complement of isSigned... nonnumeric types
2201 return false as they do with isSigned. This returns true for UByteTy,
2202 UShortTy, UIntTy, and ULongTy. </li>
2204 <li><tt>bool isInteger() const</tt>: Equivalent to isSigned() || isUnsigned().</li>
2206 <li><tt>bool isIntegral() const</tt>: Returns true if this is an integral
2207 type, which is either Bool type or one of the Integer types.</li>
2209 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
2210 floating point types.</li>
2212 <li><tt>isLosslesslyConvertableTo (const Type *Ty) const</tt>: Return true if
2213 this type can be converted to 'Ty' without any reinterpretation of bits. For
2214 example, uint to int or one pointer type to another.</li>
2218 <!-- _______________________________________________________________________ -->
2219 <div class="doc_subsubsection">
2220 <a name="m_Value">Important Derived Types</a>
2222 <div class="doc_text">
2224 <li>SequentialType : This is subclassed by ArrayType and PointerType
2226 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
2227 of the elements in the sequential type. </li>
2230 <li>ArrayType : This is a subclass of SequentialType and defines interface for
2233 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
2234 elements in the array. </li>
2237 <li>PointerType : Subclass of SequentialType for pointer types. </li>
2238 <li>StructType : subclass of DerivedTypes for struct types </li>
2239 <li>FunctionType : subclass of DerivedTypes for function types.
2241 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
2243 <li><tt> const Type * getReturnType() const</tt>: Returns the
2244 return type of the function.</li>
2245 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
2246 the type of the ith parameter.</li>
2247 <li><tt> const unsigned getNumParams() const</tt>: Returns the
2248 number of formal parameters.</li>
2254 <!-- ======================================================================= -->
2255 <div class="doc_subsection">
2256 <a name="Argument">The <tt>Argument</tt> class</a>
2259 <div class="doc_text">
2261 <p>This subclass of Value defines the interface for incoming formal
2262 arguments to a function. A Function maintains a list of its formal
2263 arguments. An argument has a pointer to the parent Function.</p>
2267 <!-- *********************************************************************** -->
2270 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
2271 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
2272 <a href="http://validator.w3.org/check/referer"><img
2273 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
2275 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
2276 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
2277 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
2278 Last modified: $Date$