1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <title>LLVM Programmer's Manual</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
10 <div class="doc_title">
11 LLVM Programmer's Manual
15 <li><a href="#introduction">Introduction</a></li>
16 <li><a href="#general">General Information</a>
18 <li><a href="#stl">The C++ Standard Template Library</a><!--
19 <li>The <tt>-time-passes</tt> option
20 <li>How to use the LLVM Makefile system
21 <li>How to write a regression test
25 <li><a href="#apis">Important and useful LLVM APIs</a>
27 <li><a href="#isa">The <tt>isa<></tt>, <tt>cast<></tt>
28 and <tt>dyn_cast<></tt> templates</a> </li>
29 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt>
32 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
33 and the <tt>-debug-only</tt> option</a> </li>
36 <li><a href="#Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
38 <li>The <tt>InstVisitor</tt> template
39 <li>The general graph API
43 <li><a href="#common">Helpful Hints for Common Operations</a>
45 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
47 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
48 in a <tt>Function</tt></a> </li>
49 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
50 in a <tt>BasicBlock</tt></a> </li>
51 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
52 in a <tt>Function</tt></a> </li>
53 <li><a href="#iterate_convert">Turning an iterator into a
54 class pointer</a> </li>
55 <li><a href="#iterate_complex">Finding call sites: a more
56 complex example</a> </li>
57 <li><a href="#calls_and_invokes">Treating calls and invokes
58 the same way</a> </li>
59 <li><a href="#iterate_chains">Iterating over def-use &
60 use-def chains</a> </li>
63 <li><a href="#simplechanges">Making simple changes</a>
65 <li><a href="#schanges_creating">Creating and inserting new
66 <tt>Instruction</tt>s</a> </li>
67 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
68 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
69 with another <tt>Value</tt></a> </li>
72 <li>Working with the Control Flow Graph
74 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
81 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
83 <li><a href="#Value">The <tt>Value</tt> class</a>
85 <li><a href="#User">The <tt>User</tt> class</a>
87 <li><a href="#Instruction">The <tt>Instruction</tt> class</a>
89 <li><a href="#GetElementPtrInst">The <tt>GetElementPtrInst</tt>
92 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
94 <li><a href="#BasicBlock">The <tt>BasicBlock</tt>class</a></li>
95 <li><a href="#Function">The <tt>Function</tt> class</a></li>
96 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class
99 <li><a href="#Module">The <tt>Module</tt> class</a></li>
100 <li><a href="#Constant">The <tt>Constant</tt> class</a></li>
101 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
102 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
105 <li><a href="#SymbolTable">The <tt>SymbolTable</tt> class </a></li>
106 <li>The <tt>ilist</tt> and <tt>iplist</tt> classes
108 <li>Creating, inserting, moving and deleting from LLVM lists </li>
111 <li>Important iterator invalidation semantics to be aware of.</li>
115 <div class="doc_author">
116 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
117 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>, and
118 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a></p>
121 <!-- *********************************************************************** -->
122 <div class="doc_section">
123 <a name="introduction">Introduction </a>
125 <!-- *********************************************************************** -->
127 <div class="doc_text">
129 <p>This document is meant to highlight some of the important classes and
130 interfaces available in the LLVM source-base. This manual is not
131 intended to explain what LLVM is, how it works, and what LLVM code looks
132 like. It assumes that you know the basics of LLVM and are interested
133 in writing transformations or otherwise analyzing or manipulating the
136 <p>This document should get you oriented so that you can find your
137 way in the continuously growing source code that makes up the LLVM
138 infrastructure. Note that this manual is not intended to serve as a
139 replacement for reading the source code, so if you think there should be
140 a method in one of these classes to do something, but it's not listed,
141 check the source. Links to the <a href="/doxygen/">doxygen</a> sources
142 are provided to make this as easy as possible.</p>
144 <p>The first section of this document describes general information that is
145 useful to know when working in the LLVM infrastructure, and the second describes
146 the Core LLVM classes. In the future this manual will be extended with
147 information describing how to use extension libraries, such as dominator
148 information, CFG traversal routines, and useful utilities like the <tt><a
149 href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
153 <!-- *********************************************************************** -->
154 <div class="doc_section">
155 <a name="general">General Information</a>
157 <!-- *********************************************************************** -->
159 <div class="doc_text">
161 <p>This section contains general information that is useful if you are working
162 in the LLVM source-base, but that isn't specific to any particular API.</p>
166 <!-- ======================================================================= -->
167 <div class="doc_subsection">
168 <a name="stl">The C++ Standard Template Library</a>
171 <div class="doc_text">
173 <p>LLVM makes heavy use of the C++ Standard Template Library (STL),
174 perhaps much more than you are used to, or have seen before. Because of
175 this, you might want to do a little background reading in the
176 techniques used and capabilities of the library. There are many good
177 pages that discuss the STL, and several books on the subject that you
178 can get, so it will not be discussed in this document.</p>
180 <p>Here are some useful links:</p>
184 <li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
185 reference</a> - an excellent reference for the STL and other parts of the
186 standard C++ library.</li>
188 <li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
189 O'Reilly book in the making. It has a decent <a
190 href="http://www.tempest-sw.com/cpp/ch13-libref.html">Standard Library
191 Reference</a> that rivals Dinkumware's, and is actually free until the book is
194 <li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
197 <li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
199 href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
202 <li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
205 <li><a href="http://www.linux.com.cn/Bruce_Eckel/TICPPv2/Contents.htm">
206 Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
211 <p>You are also encouraged to take a look at the <a
212 href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
213 to write maintainable code more than where to put your curly braces.</p>
217 <!-- ======================================================================= -->
218 <div class="doc_subsection">
219 <a name="stl">Other useful references</a>
222 <div class="doc_text">
224 <p>LLVM is currently using CVS as its source versioning system. You may find
225 this reference handy:</p>
228 <li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
229 Branch and Tag Primer</a></li>
234 <!-- *********************************************************************** -->
235 <div class="doc_section">
236 <a name="apis">Important and useful LLVM APIs</a>
238 <!-- *********************************************************************** -->
240 <div class="doc_text">
242 <p>Here we highlight some LLVM APIs that are generally useful and good to
243 know about when writing transformations.</p>
247 <!-- ======================================================================= -->
248 <div class="doc_subsection">
249 <a name="isa">The isa<>, cast<> and dyn_cast<> templates</a>
252 <div class="doc_text">
254 <p>The LLVM source-base makes extensive use of a custom form of RTTI.
255 These templates have many similarities to the C++ <tt>dynamic_cast<></tt>
256 operator, but they don't have some drawbacks (primarily stemming from
257 the fact that <tt>dynamic_cast<></tt> only works on classes that
258 have a v-table). Because they are used so often, you must know what they
259 do and how they work. All of these templates are defined in the <a
260 href="/doxygen/Casting_8h-source.html"><tt>Support/Casting.h</tt></a>
261 file (note that you very rarely have to include this file directly).</p>
264 <dt><tt>isa<></tt>: </dt>
266 <dd>The <tt>isa<></tt> operator works exactly like the Java
267 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
268 a reference or pointer points to an instance of the specified class. This can
269 be very useful for constraint checking of various sorts (example below).</dd>
271 <dt><tt>cast<></tt>: </dt>
273 <dd>The <tt>cast<></tt> operator is a "checked cast" operation. It
274 converts a pointer or reference from a base class to a derived cast, causing
275 an assertion failure if it is not really an instance of the right type. This
276 should be used in cases where you have some information that makes you believe
277 that something is of the right type. An example of the <tt>isa<></tt>
278 and <tt>cast<></tt> template is:
281 static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
282 if (isa<<a href="#Constant">Constant</a>>(V) || isa<<a href="#Argument">Argument</a>>(V) || isa<<a href="#GlobalValue">GlobalValue</a>>(V))
285 <i>// Otherwise, it must be an instruction...</i>
286 return !L->contains(cast<<a href="#Instruction">Instruction</a>>(V)->getParent());
289 <p>Note that you should <b>not</b> use an <tt>isa<></tt> test followed
290 by a <tt>cast<></tt>, for that use the <tt>dyn_cast<></tt>
295 <dt><tt>dyn_cast<></tt>:</dt>
297 <dd>The <tt>dyn_cast<></tt> operator is a "checking cast" operation. It
298 checks to see if the operand is of the specified type, and if so, returns a
299 pointer to it (this operator does not work with references). If the operand is
300 not of the correct type, a null pointer is returned. Thus, this works very
301 much like the <tt>dynamic_cast</tt> operator in C++, and should be used in the
302 same circumstances. Typically, the <tt>dyn_cast<></tt> operator is used
303 in an <tt>if</tt> statement or some other flow control statement like this:
306 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast<<a href="#AllocationInst">AllocationInst</a>>(Val)) {
311 <p> This form of the <tt>if</tt> statement effectively combines together a
312 call to <tt>isa<></tt> and a call to <tt>cast<></tt> into one
313 statement, which is very convenient.</p>
315 <p> Another common example is:</p>
318 <i>// Loop over all of the phi nodes in a basic block</i>
319 BasicBlock::iterator BBI = BB->begin();
320 for (; <a href="#PhiNode">PHINode</a> *PN = dyn_cast<<a href="#PHINode">PHINode</a>>(BBI); ++BBI)
321 std::cerr << *PN;
324 <p>Note that the <tt>dyn_cast<></tt> operator, like C++'s
325 <tt>dynamic_cast</tt> or Java's <tt>instanceof</tt> operator, can be abused.
326 In particular you should not use big chained <tt>if/then/else</tt> blocks to
327 check for lots of different variants of classes. If you find yourself
328 wanting to do this, it is much cleaner and more efficient to use the
329 InstVisitor class to dispatch over the instruction type directly.</p>
333 <dt><tt>cast_or_null<></tt>: </dt>
335 <dd>The <tt>cast_or_null<></tt> operator works just like the
336 <tt>cast<></tt> operator, except that it allows for a null pointer as
337 an argument (which it then propagates). This can sometimes be useful,
338 allowing you to combine several null checks into one.</dd>
340 <dt><tt>dyn_cast_or_null<></tt>: </dt>
342 <dd>The <tt>dyn_cast_or_null<></tt> operator works just like the
343 <tt>dyn_cast<></tt> operator, except that it allows for a null pointer
344 as an argument (which it then propagates). This can sometimes be useful,
345 allowing you to combine several null checks into one.</dd>
349 <p>These five templates can be used with any classes, whether they have a
350 v-table or not. To add support for these templates, you simply need to add
351 <tt>classof</tt> static methods to the class you are interested casting
352 to. Describing this is currently outside the scope of this document, but there
353 are lots of examples in the LLVM source base.</p>
357 <!-- ======================================================================= -->
358 <div class="doc_subsection">
359 <a name="DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt> option</a>
362 <div class="doc_text">
364 <p>Often when working on your pass you will put a bunch of debugging printouts
365 and other code into your pass. After you get it working, you want to remove
366 it... but you may need it again in the future (to work out new bugs that you run
369 <p> Naturally, because of this, you don't want to delete the debug printouts,
370 but you don't want them to always be noisy. A standard compromise is to comment
371 them out, allowing you to enable them if you need them in the future.</p>
373 <p>The "<tt><a href="/doxygen/Debug_8h-source.html">Support/Debug.h</a></tt>"
374 file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
375 this problem. Basically, you can put arbitrary code into the argument of the
376 <tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
377 tool) is run with the '<tt>-debug</tt>' command line argument:</p>
379 <pre> ... <br> DEBUG(std::cerr << "I am here!\n");<br> ...<br></pre>
381 <p>Then you can run your pass like this:</p>
383 <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>
385 <p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
386 to not have to create "yet another" command line option for the debug output for
387 your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
388 so they do not cause a performance impact at all (for the same reason, they
389 should also not contain side-effects!).</p>
391 <p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
392 enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
393 "<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
394 program hasn't been started yet, you can always just run it with
399 <!-- _______________________________________________________________________ -->
400 <div class="doc_subsubsection">
401 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE()</tt> and
402 the <tt>-debug-only</tt> option</a>
405 <div class="doc_text">
407 <p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
408 just turns on <b>too much</b> information (such as when working on the code
409 generator). If you want to enable debug information with more fine-grained
410 control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
411 option as follows:</p>
413 <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>
415 <p>Then you can run your pass like this:</p>
417 <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>
419 <p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
420 a file, to specify the debug type for the entire module (if you do this before
421 you <tt>#include "Support/Debug.h"</tt>, you don't have to insert the ugly
422 <tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
423 "bar", because there is no system in place to ensure that names do not
424 conflict. If two different modules use the same string, they will all be turned
425 on when the name is specified. This allows, for example, all debug information
426 for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
427 even if the source lives in multiple files.</p>
431 <!-- ======================================================================= -->
432 <div class="doc_subsection">
433 <a name="Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
437 <div class="doc_text">
440 href="/doxygen/Statistic_8h-source.html">Support/Statistic.h</a></tt>" file
441 provides a template named <tt>Statistic</tt> that is used as a unified way to
442 keep track of what the LLVM compiler is doing and how effective various
443 optimizations are. It is useful to see what optimizations are contributing to
444 making a particular program run faster.</p>
446 <p>Often you may run your pass on some big program, and you're interested to see
447 how many times it makes a certain transformation. Although you can do this with
448 hand inspection, or some ad-hoc method, this is a real pain and not very useful
449 for big programs. Using the <tt>Statistic</tt> template makes it very easy to
450 keep track of this information, and the calculated information is presented in a
451 uniform manner with the rest of the passes being executed.</p>
453 <p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
454 it are as follows:</p>
457 <li>Define your statistic like this:
458 <pre>static Statistic<> NumXForms("mypassname", "The # of times I did stuff");<br></pre>
460 <p>The <tt>Statistic</tt> template can emulate just about any data-type,
461 but if you do not specify a template argument, it defaults to acting like
462 an unsigned int counter (this is usually what you want).</p></li>
464 <li>Whenever you make a transformation, bump the counter:
465 <pre> ++NumXForms; // I did stuff<br></pre>
469 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
470 statistics gathered, use the '<tt>-stats</tt>' option:</p>
472 <pre> $ opt -stats -mypassname < program.bc > /dev/null<br> ... statistic output ...<br></pre>
474 <p> When running <tt>gccas</tt> on a C file from the SPEC benchmark
475 suite, it gives a report that looks like this:</p>
477 <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>
479 <p>Obviously, with so many optimizations, having a unified framework for this
480 stuff is very nice. Making your pass fit well into the framework makes it more
481 maintainable and useful.</p>
485 <!-- *********************************************************************** -->
486 <div class="doc_section">
487 <a name="common">Helpful Hints for Common Operations</a>
489 <!-- *********************************************************************** -->
491 <div class="doc_text">
493 <p>This section describes how to perform some very simple transformations of
494 LLVM code. This is meant to give examples of common idioms used, showing the
495 practical side of LLVM transformations. <p> Because this is a "how-to" section,
496 you should also read about the main classes that you will be working with. The
497 <a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
498 and descriptions of the main classes that you should know about.</p>
502 <!-- NOTE: this section should be heavy on example code -->
503 <!-- ======================================================================= -->
504 <div class="doc_subsection">
505 <a name="inspection">Basic Inspection and Traversal Routines</a>
508 <div class="doc_text">
510 <p>The LLVM compiler infrastructure have many different data structures that may
511 be traversed. Following the example of the C++ standard template library, the
512 techniques used to traverse these various data structures are all basically the
513 same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
514 method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
515 function returns an iterator pointing to one past the last valid element of the
516 sequence, and there is some <tt>XXXiterator</tt> data type that is common
517 between the two operations.</p>
519 <p>Because the pattern for iteration is common across many different aspects of
520 the program representation, the standard template library algorithms may be used
521 on them, and it is easier to remember how to iterate. First we show a few common
522 examples of the data structures that need to be traversed. Other data
523 structures are traversed in very similar ways.</p>
527 <!-- _______________________________________________________________________ -->
528 <div class="doc_subsubsection">
529 <a name="iterate_function">Iterating over the </a><a
530 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
531 href="#Function"><tt>Function</tt></a>
534 <div class="doc_text">
536 <p>It's quite common to have a <tt>Function</tt> instance that you'd like to
537 transform in some way; in particular, you'd like to manipulate its
538 <tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
539 the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
540 an example that prints the name of a <tt>BasicBlock</tt> and the number of
541 <tt>Instruction</tt>s it contains:</p>
543 <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>
545 <p>Note that i can be used as if it were a pointer for the purposes of
546 invoking member functions of the <tt>Instruction</tt> class. This is
547 because the indirection operator is overloaded for the iterator
548 classes. In the above code, the expression <tt>i->size()</tt> is
549 exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
553 <!-- _______________________________________________________________________ -->
554 <div class="doc_subsubsection">
555 <a name="iterate_basicblock">Iterating over the </a><a
556 href="#Instruction"><tt>Instruction</tt></a>s in a <a
557 href="#BasicBlock"><tt>BasicBlock</tt></a>
560 <div class="doc_text">
562 <p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
563 easy to iterate over the individual instructions that make up
564 <tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
565 a <tt>BasicBlock</tt>:</p>
567 <pre> // blk is a pointer to a BasicBlock instance<br> for (BasicBlock::iterator i = blk->begin(), e = blk->end(); i != e; ++i)<br> // the next statement works since operator<<(ostream&,...) <br> // is overloaded for Instruction&<br> cerr << *i << "\n";<br></pre>
569 <p>However, this isn't really the best way to print out the contents of a
570 <tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
571 anything you'll care about, you could have just invoked the print routine on the
572 basic block itself: <tt>cerr << *blk << "\n";</tt>.</p>
574 <p>Note that currently operator<< is implemented for <tt>Value*</tt>, so
575 it will print out the contents of the pointer, instead of the pointer value you
576 might expect. This is a deprecated interface that will be removed in the
577 future, so it's best not to depend on it. To print out the pointer value for
578 now, you must cast to <tt>void*</tt>.</p>
582 <!-- _______________________________________________________________________ -->
583 <div class="doc_subsubsection">
584 <a name="iterate_institer">Iterating over the </a><a
585 href="#Instruction"><tt>Instruction</tt></a>s in a <a
586 href="#Function"><tt>Function</tt></a>
589 <div class="doc_text">
591 <p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
592 <tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
593 <tt>InstIterator</tt> should be used instead. You'll need to include <a
594 href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
595 and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
596 small example that shows how to dump all instructions in a function to the standard error stream:<p>
598 <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>
599 Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
600 worklist with its initial contents. For example, if you wanted to
601 initialize a worklist to contain all instructions in a <tt>Function</tt>
602 F, all you would need to do is something like:
603 <pre>std::set<Instruction*> worklist;<br>worklist.insert(inst_begin(F), inst_end(F));<br></pre>
605 <p>The STL set <tt>worklist</tt> would now contain all instructions in the
606 <tt>Function</tt> pointed to by F.</p>
610 <!-- _______________________________________________________________________ -->
611 <div class="doc_subsubsection">
612 <a name="iterate_convert">Turning an iterator into a class pointer (and
616 <div class="doc_text">
618 <p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
619 instance when all you've got at hand is an iterator. Well, extracting
620 a reference or a pointer from an iterator is very straight-forward.
621 Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
622 is a <tt>BasicBlock::const_iterator</tt>:</p>
624 <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>
626 <p>However, the iterators you'll be working with in the LLVM framework are
627 special: they will automatically convert to a ptr-to-instance type whenever they
628 need to. Instead of dereferencing the iterator and then taking the address of
629 the result, you can simply assign the iterator to the proper pointer type and
630 you get the dereference and address-of operation as a result of the assignment
631 (behind the scenes, this is a result of overloading casting mechanisms). Thus
632 the last line of the last example,</p>
634 <pre>Instruction* pinst = &*i;</pre>
636 <p>is semantically equivalent to</p>
638 <pre>Instruction* pinst = i;</pre>
640 <p>It's also possible to turn a class pointer into the corresponding iterator,
641 and this is a constant time operation (very efficient). The following code
642 snippet illustrates use of the conversion constructors provided by LLVM
643 iterators. By using these, you can explicitly grab the iterator of something
644 without actually obtaining it via iteration over some structure:</p>
646 <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>
650 <!--_______________________________________________________________________-->
651 <div class="doc_subsubsection">
652 <a name="iterate_complex">Finding call sites: a slightly more complex
656 <div class="doc_text">
658 <p>Say that you're writing a FunctionPass and would like to count all the
659 locations in the entire module (that is, across every <tt>Function</tt>) where a
660 certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
661 learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
662 much more straight-forward manner, but this example will allow us to explore how
663 you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudocode, this
664 is what we want to do:</p>
666 <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>
668 <p>And the actual code is (remember, since we're writing a
669 <tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
670 override the <tt>runOnFunction</tt> method...):</p>
672 <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
673 href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a><<a
674 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>
678 <!--_______________________________________________________________________-->
679 <div class="doc_subsubsection">
680 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
683 <div class="doc_text">
685 <p>You may have noticed that the previous example was a bit oversimplified in
686 that it did not deal with call sites generated by 'invoke' instructions. In
687 this, and in other situations, you may find that you want to treat
688 <tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
689 most-specific common base class is <tt>Instruction</tt>, which includes lots of
690 less closely-related things. For these cases, LLVM provides a handy wrapper
692 href="http://llvm.cs.uiuc.edu/doxygen/classCallSite.html"><tt>CallSite</tt></a>.
693 It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
694 methods that provide functionality common to <tt>CallInst</tt>s and
695 <tt>InvokeInst</tt>s.</p>
697 <p>This class has "value semantics": it should be passed by value, not by
698 reference and it should not be dynamically allocated or deallocated using
699 <tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
700 assignable and constructable, with costs equivalents to that of a bare pointer.
701 If you look at its definition, it has only a single pointer member.</p>
705 <!--_______________________________________________________________________-->
706 <div class="doc_subsubsection">
707 <a name="iterate_chains">Iterating over def-use & use-def chains</a>
710 <div class="doc_text">
712 <p>Frequently, we might have an instance of the <a
713 href="/doxygen/classValue.html">Value Class</a> and we want to determine which
714 <tt>User</tt>s use the <tt>Value</tt>. The list of all <tt>User</tt>s of a
715 particular <tt>Value</tt> is called a <i>def-use</i> chain. For example, let's
716 say we have a <tt>Function*</tt> named <tt>F</tt> to a particular function
717 <tt>foo</tt>. Finding all of the instructions that <i>use</i> <tt>foo</tt> is as
718 simple as iterating over the <i>def-use</i> chain of <tt>F</tt>:</p>
720 <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>
722 <p>Alternately, it's common to have an instance of the <a
723 href="/doxygen/classUser.html">User Class</a> and need to know what
724 <tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
725 <tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
726 <tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
727 all of the values that a particular instruction uses (that is, the operands of
728 the particular <tt>Instruction</tt>):</p>
730 <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>
733 def-use chains ("finding all users of"): Value::use_begin/use_end
734 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
739 <!-- ======================================================================= -->
740 <div class="doc_subsection">
741 <a name="simplechanges">Making simple changes</a>
744 <div class="doc_text">
746 <p>There are some primitive transformation operations present in the LLVM
747 infrastructure that are worth knowing about. When performing
748 transformations, it's fairly common to manipulate the contents of basic
749 blocks. This section describes some of the common methods for doing so
750 and gives example code.</p>
754 <!--_______________________________________________________________________-->
755 <div class="doc_subsubsection">
756 <a name="schanges_creating">Creating and inserting new
757 <tt>Instruction</tt>s</a>
760 <div class="doc_text">
762 <p><i>Instantiating Instructions</i></p>
764 <p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
765 constructor for the kind of instruction to instantiate and provide the necessary
766 parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
767 (const-ptr-to) <tt>Type</tt>. Thus:</p>
769 <pre>AllocaInst* ai = new AllocaInst(Type::IntTy);</pre>
771 <p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
772 one integer in the current stack frame, at runtime. Each <tt>Instruction</tt>
773 subclass is likely to have varying default parameters which change the semantics
774 of the instruction, so refer to the <a
775 href="/doxygen/classInstruction.html">doxygen documentation for the subclass of
776 Instruction</a> that you're interested in instantiating.</p>
778 <p><i>Naming values</i></p>
780 <p>It is very useful to name the values of instructions when you're able to, as
781 this facilitates the debugging of your transformations. If you end up looking
782 at generated LLVM machine code, you definitely want to have logical names
783 associated with the results of instructions! By supplying a value for the
784 <tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
785 associate a logical name with the result of the instruction's execution at
786 runtime. For example, say that I'm writing a transformation that dynamically
787 allocates space for an integer on the stack, and that integer is going to be
788 used as some kind of index by some other code. To accomplish this, I place an
789 <tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
790 <tt>Function</tt>, and I'm intending to use it within the same
791 <tt>Function</tt>. I might do:</p>
793 <pre>AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");</pre>
795 <p>where <tt>indexLoc</tt> is now the logical name of the instruction's
796 execution value, which is a pointer to an integer on the runtime stack.</p>
798 <p><i>Inserting instructions</i></p>
800 <p>There are essentially two ways to insert an <tt>Instruction</tt>
801 into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
804 <li>Insertion into an explicit instruction list
806 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
807 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
808 before <tt>*pi</tt>, we do the following: </p>
810 <pre> BasicBlock *pb = ...;<br> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pb->getInstList().insert(pi, newInst); // inserts newInst before pi in pb<br></pre></li>
812 <li>Insertion into an implicit instruction list
814 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
815 are implicitly associated with an existing instruction list: the instruction
816 list of the enclosing basic block. Thus, we could have accomplished the same
817 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
820 <pre> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pi->getParent()->getInstList().insert(pi, newInst);<br></pre>
822 <p>In fact, this sequence of steps occurs so frequently that the
823 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
824 constructors which take (as a default parameter) a pointer to an
825 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
826 precede. That is, <tt>Instruction</tt> constructors are capable of
827 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
828 provided instruction, immediately before that instruction. Using an
829 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
830 parameter, the above code becomes:</p>
832 <pre>Instruction* pi = ...;<br>Instruction* newInst = new Instruction(..., pi);<br></pre>
834 <p>which is much cleaner, especially if you're creating a lot of
835 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
840 <!--_______________________________________________________________________-->
841 <div class="doc_subsubsection">
842 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
845 <div class="doc_text">
847 <p>Deleting an instruction from an existing sequence of instructions that form a
848 <a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
849 you must have a pointer to the instruction that you wish to delete. Second, you
850 need to obtain the pointer to that instruction's basic block. You use the
851 pointer to the basic block to get its list of instructions and then use the
852 erase function to remove your instruction. For example:</p>
854 <pre> <a href="#Instruction">Instruction</a> *I = .. ;<br> <a
855 href="#BasicBlock">BasicBlock</a> *BB = I->getParent();<br> BB->getInstList().erase(I);<br></pre>
859 <!--_______________________________________________________________________-->
860 <div class="doc_subsubsection">
861 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
865 <div class="doc_text">
867 <p><i>Replacing individual instructions</i></p>
869 <p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
870 permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
871 and <tt>ReplaceInstWithInst</tt>.</p>
873 <h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
876 <li><tt>ReplaceInstWithValue</tt>
878 <p>This function replaces all uses (within a basic block) of a given
879 instruction with a value, and then removes the original instruction. The
880 following example illustrates the replacement of the result of a particular
881 <tt>AllocaInst</tt> that allocates memory for a single integer with an null
882 pointer to an integer.</p>
884 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithValue(instToReplace->getParent()->getInstList(), ii,<br> Constant::getNullValue(PointerType::get(Type::IntTy)));<br></pre></li>
886 <li><tt>ReplaceInstWithInst</tt>
888 <p>This function replaces a particular instruction with another
889 instruction. The following example illustrates the replacement of one
890 <tt>AllocaInst</tt> with another.</p>
892 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithInst(instToReplace->getParent()->getInstList(), ii,<br> new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));<br></pre></li>
895 <p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
897 <p>You can use <tt>Value::replaceAllUsesWith</tt> and
898 <tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
899 doxygen documentation for the <a href="/doxygen/classValue.html">Value Class</a>
900 and <a href="/doxygen/classUser.html">User Class</a>, respectively, for more
903 <!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
904 include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
905 ReplaceInstWithValue, ReplaceInstWithInst -->
909 <!-- *********************************************************************** -->
910 <div class="doc_section">
911 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
913 <!-- *********************************************************************** -->
915 <div class="doc_text">
917 <p>The Core LLVM classes are the primary means of representing the program
918 being inspected or transformed. The core LLVM classes are defined in
919 header files in the <tt>include/llvm/</tt> directory, and implemented in
920 the <tt>lib/VMCore</tt> directory.</p>
924 <!-- ======================================================================= -->
925 <div class="doc_subsection">
926 <a name="Value">The <tt>Value</tt> class</a>
931 <p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
933 doxygen info: <a href="/doxygen/classValue.html">Value Class</a></p>
935 <p>The <tt>Value</tt> class is the most important class in the LLVM Source
936 base. It represents a typed value that may be used (among other things) as an
937 operand to an instruction. There are many different types of <tt>Value</tt>s,
938 such as <a href="#Constant"><tt>Constant</tt></a>s,<a
939 href="#Argument"><tt>Argument</tt></a>s. Even <a
940 href="#Instruction"><tt>Instruction</tt></a>s and <a
941 href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
943 <p>A particular <tt>Value</tt> may be used many times in the LLVM representation
944 for a program. For example, an incoming argument to a function (represented
945 with an instance of the <a href="#Argument">Argument</a> class) is "used" by
946 every instruction in the function that references the argument. To keep track
947 of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
948 href="#User"><tt>User</tt></a>s that is using it (the <a
949 href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
950 graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
951 def-use information in the program, and is accessible through the <tt>use_</tt>*
952 methods, shown below.</p>
954 <p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
955 and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
956 method. In addition, all LLVM values can be named. The "name" of the
957 <tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
959 <pre> %<b>foo</b> = add int 1, 2<br></pre>
961 <p><a name="#nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
962 that the name of any value may be missing (an empty string), so names should
963 <b>ONLY</b> be used for debugging (making the source code easier to read,
964 debugging printouts), they should not be used to keep track of values or map
965 between them. For this purpose, use a <tt>std::map</tt> of pointers to the
966 <tt>Value</tt> itself instead.</p>
968 <p>One important aspect of LLVM is that there is no distinction between an SSA
969 variable and the operation that produces it. Because of this, any reference to
970 the value produced by an instruction (or the value available as an incoming
971 argument, for example) is represented as a direct pointer to the instance of
973 represents this value. Although this may take some getting used to, it
974 simplifies the representation and makes it easier to manipulate.</p>
978 <!-- _______________________________________________________________________ -->
979 <div class="doc_subsubsection">
980 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
983 <div class="doc_text">
986 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
988 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
990 <tt>unsigned use_size()</tt> - Returns the number of users of the
992 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
993 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
995 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
997 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
999 <p> These methods are the interface to access the def-use
1000 information in LLVM. As with all other iterators in LLVM, the naming
1001 conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
1003 <li><tt><a href="#Type">Type</a> *getType() const</tt>
1004 <p>This method returns the Type of the Value.</p>
1006 <li><tt>bool hasName() const</tt><br>
1007 <tt>std::string getName() const</tt><br>
1008 <tt>void setName(const std::string &Name)</tt>
1009 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
1010 be aware of the <a href="#nameWarning">precaution above</a>.</p>
1012 <li><tt>void replaceAllUsesWith(Value *V)</tt>
1014 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
1015 href="#User"><tt>User</tt>s</a> of the current value to refer to
1016 "<tt>V</tt>" instead. For example, if you detect that an instruction always
1017 produces a constant value (for example through constant folding), you can
1018 replace all uses of the instruction with the constant like this:</p>
1020 <pre> Inst->replaceAllUsesWith(ConstVal);<br></pre>
1025 <!-- ======================================================================= -->
1026 <div class="doc_subsection">
1027 <a name="User">The <tt>User</tt> class</a>
1030 <div class="doc_text">
1033 <tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
1034 doxygen info: <a href="/doxygen/classUser.html">User Class</a><br>
1035 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1037 <p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
1038 refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
1039 that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
1040 referring to. The <tt>User</tt> class itself is a subclass of
1043 <p>The operands of a <tt>User</tt> point directly to the LLVM <a
1044 href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
1045 Single Assignment (SSA) form, there can only be one definition referred to,
1046 allowing this direct connection. This connection provides the use-def
1047 information in LLVM.</p>
1051 <!-- _______________________________________________________________________ -->
1052 <div class="doc_subsubsection">
1053 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
1056 <div class="doc_text">
1058 <p>The <tt>User</tt> class exposes the operand list in two ways: through
1059 an index access interface and through an iterator based interface.</p>
1062 <li><tt>Value *getOperand(unsigned i)</tt><br>
1063 <tt>unsigned getNumOperands()</tt>
1064 <p> These two methods expose the operands of the <tt>User</tt> in a
1065 convenient form for direct access.</p></li>
1067 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
1069 <tt>User::op_const_iterator</tt> <tt>use_iterator op_begin()</tt> -
1070 Get an iterator to the start of the operand list.<br>
1071 <tt>use_iterator op_end()</tt> - Get an iterator to the end of the
1073 <p> Together, these methods make up the iterator based interface to
1074 the operands of a <tt>User</tt>.</p></li>
1079 <!-- ======================================================================= -->
1080 <div class="doc_subsection">
1081 <a name="Instruction">The <tt>Instruction</tt> class</a>
1084 <div class="doc_text">
1086 <p><tt>#include "</tt><tt><a
1087 href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
1088 doxygen info: <a href="/doxygen/classInstruction.html">Instruction Class</a><br>
1089 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1090 href="#Value"><tt>Value</tt></a></p>
1092 <p>The <tt>Instruction</tt> class is the common base class for all LLVM
1093 instructions. It provides only a few methods, but is a very commonly used
1094 class. The primary data tracked by the <tt>Instruction</tt> class itself is the
1095 opcode (instruction type) and the parent <a
1096 href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
1097 into. To represent a specific type of instruction, one of many subclasses of
1098 <tt>Instruction</tt> are used.</p>
1100 <p> Because the <tt>Instruction</tt> class subclasses the <a
1101 href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
1102 way as for other <a href="#User"><tt>User</tt></a>s (with the
1103 <tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
1104 <tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
1105 the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
1106 file contains some meta-data about the various different types of instructions
1107 in LLVM. It describes the enum values that are used as opcodes (for example
1108 <tt>Instruction::Add</tt> and <tt>Instruction::SetLE</tt>), as well as the
1109 concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
1110 example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
1111 href="#SetCondInst">SetCondInst</a></tt>). Unfortunately, the use of macros in
1112 this file confuses doxygen, so these enum values don't show up correctly in the
1113 <a href="/doxygen/classInstruction.html">doxygen output</a>.</p>
1117 <!-- _______________________________________________________________________ -->
1118 <div class="doc_subsubsection">
1119 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
1123 <div class="doc_text">
1126 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
1127 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
1128 this <tt>Instruction</tt> is embedded into.</p></li>
1129 <li><tt>bool mayWriteToMemory()</tt>
1130 <p>Returns true if the instruction writes to memory, i.e. it is a
1131 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
1132 <li><tt>unsigned getOpcode()</tt>
1133 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
1134 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
1135 <p>Returns another instance of the specified instruction, identical
1136 in all ways to the original except that the instruction has no parent
1137 (ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
1138 and it has no name</p></li>
1143 <!-- ======================================================================= -->
1144 <div class="doc_subsection">
1145 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
1148 <div class="doc_text">
1150 <p><tt>#include "<a href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
1151 doxygen info: <a href="/doxygen/classBasicBlock.html">BasicBlock Class</a><br>
1152 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1154 <p>This class represents a single entry multiple exit section of the code,
1155 commonly known as a basic block by the compiler community. The
1156 <tt>BasicBlock</tt> class maintains a list of <a
1157 href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
1158 Matching the language definition, the last element of this list of instructions
1159 is always a terminator instruction (a subclass of the <a
1160 href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
1162 <p>In addition to tracking the list of instructions that make up the block, the
1163 <tt>BasicBlock</tt> class also keeps track of the <a
1164 href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
1166 <p>Note that <tt>BasicBlock</tt>s themselves are <a
1167 href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
1168 like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
1173 <!-- _______________________________________________________________________ -->
1174 <div class="doc_subsubsection">
1175 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
1179 <div class="doc_text">
1182 <li><tt>BasicBlock(const std::string &Name = "", </tt><tt><a
1183 href="#Function">Function</a> *Parent = 0)</tt>
1184 <p>The <tt>BasicBlock</tt> constructor is used to create new basic
1185 blocks for insertion into a function. The constructor optionally takes
1186 a name for the new block, and a <a href="#Function"><tt>Function</tt></a>
1187 to insert it into. If the <tt>Parent</tt> parameter is specified, the
1188 new <tt>BasicBlock</tt> is automatically inserted at the end of the
1189 specified <a href="#Function"><tt>Function</tt></a>, if not specified,
1190 the BasicBlock must be manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p>
1192 <li><tt>BasicBlock::iterator</tt> - Typedef for instruction list
1194 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
1195 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,<tt>size()</tt>,<tt>empty()</tt>,<tt>rbegin()</tt>,<tt>rend()
1196 - </tt>STL style functions for accessing the instruction list.
1197 <p> These methods and typedefs are forwarding functions that have
1198 the same semantics as the standard library methods of the same names.
1199 These methods expose the underlying instruction list of a basic block in
1200 a way that is easy to manipulate. To get the full complement of
1201 container operations (including operations to update the list), you must
1202 use the <tt>getInstList()</tt> method.</p></li>
1203 <li><tt>BasicBlock::InstListType &getInstList()</tt>
1204 <p> This method is used to get access to the underlying container
1205 that actually holds the Instructions. This method must be used when
1206 there isn't a forwarding function in the <tt>BasicBlock</tt> class for
1207 the operation that you would like to perform. Because there are no
1208 forwarding functions for "updating" operations, you need to use this if
1209 you want to update the contents of a <tt>BasicBlock</tt>.</p></li>
1210 <li><tt><a href="#Function">Function</a> *getParent()</tt>
1211 <p> Returns a pointer to <a href="#Function"><tt>Function</tt></a>
1212 the block is embedded into, or a null pointer if it is homeless.</p></li>
1213 <li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
1214 <p> Returns a pointer to the terminator instruction that appears at
1215 the end of the <tt>BasicBlock</tt>. If there is no terminator
1216 instruction, or if the last instruction in the block is not a
1217 terminator, then a null pointer is returned.</p></li>
1222 <!-- ======================================================================= -->
1223 <div class="doc_subsection">
1224 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
1227 <div class="doc_text">
1230 href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
1231 doxygen info: <a href="/doxygen/classGlobalValue.html">GlobalValue Class</a><br>
1232 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1233 href="#Value"><tt>Value</tt></a></p>
1235 <p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
1236 href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
1237 visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
1238 Because they are visible at global scope, they are also subject to linking with
1239 other globals defined in different translation units. To control the linking
1240 process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
1241 <tt>GlobalValue</tt>s know whether they have internal or external linkage, as
1242 defined by the <tt>LinkageTypes</tt> enumerator.</p>
1244 <p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
1245 <tt>static</tt> in C), it is not visible to code outside the current translation
1246 unit, and does not participate in linking. If it has external linkage, it is
1247 visible to external code, and does participate in linking. In addition to
1248 linkage information, <tt>GlobalValue</tt>s keep track of which <a
1249 href="#Module"><tt>Module</tt></a> they are currently part of.</p>
1251 <p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
1252 by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
1253 global is always a pointer to its contents. It is important to remember this
1254 when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
1255 be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
1256 subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
1257 int]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
1258 the address of the first element of this array and the value of the
1259 <tt>GlobalVariable</tt> are the same, they have different types. The
1260 <tt>GlobalVariable</tt>'s type is <tt>[24 x int]</tt>. The first element's type
1261 is <tt>int.</tt> Because of this, accessing a global value requires you to
1262 dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
1263 can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
1264 Language Reference Manual</a>.</p>
1268 <!-- _______________________________________________________________________ -->
1269 <div class="doc_subsubsection">
1270 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
1274 <div class="doc_text">
1277 <li><tt>bool hasInternalLinkage() const</tt><br>
1278 <tt>bool hasExternalLinkage() const</tt><br>
1279 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
1280 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
1283 <li><tt><a href="#Module">Module</a> *getParent()</tt>
1284 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
1285 GlobalValue is currently embedded into.</p></li>
1290 <!-- ======================================================================= -->
1291 <div class="doc_subsection">
1292 <a name="Function">The <tt>Function</tt> class</a>
1295 <div class="doc_text">
1298 href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
1299 info: <a href="/doxygen/classFunction.html">Function Class</a><br> Superclasses:
1300 <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1301 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1303 <p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
1304 actually one of the more complex classes in the LLVM heirarchy because it must
1305 keep track of a large amount of data. The <tt>Function</tt> class keeps track
1306 of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal <a
1307 href="#Argument"><tt>Argument</tt></a>s, and a <a
1308 href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
1310 <p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
1311 commonly used part of <tt>Function</tt> objects. The list imposes an implicit
1312 ordering of the blocks in the function, which indicate how the code will be
1313 layed out by the backend. Additionally, the first <a
1314 href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
1315 <tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
1316 block. There are no implicit exit nodes, and in fact there may be multiple exit
1317 nodes from a single <tt>Function</tt>. If the <a
1318 href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
1319 the <tt>Function</tt> is actually a function declaration: the actual body of the
1320 function hasn't been linked in yet.</p>
1322 <p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
1323 <tt>Function</tt> class also keeps track of the list of formal <a
1324 href="#Argument"><tt>Argument</tt></a>s that the function receives. This
1325 container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
1326 nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
1327 the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
1329 <p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
1330 LLVM feature that is only used when you have to look up a value by name. Aside
1331 from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
1332 internally to make sure that there are not conflicts between the names of <a
1333 href="#Instruction"><tt>Instruction</tt></a>s, <a
1334 href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
1335 href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
1339 <!-- _______________________________________________________________________ -->
1340 <div class="doc_subsubsection">
1341 <a name="m_Function">Important Public Members of the <tt>Function</tt>
1345 <div class="doc_text">
1348 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
1349 *Ty, bool isInternal, const std::string &N = "", Module* Parent = 0)</tt>
1351 <p>Constructor used when you need to create new <tt>Function</tt>s to add
1352 the the program. The constructor must specify the type of the function to
1353 create and whether or not it should start out with internal or external
1354 linkage. The <a href="#FunctionType"><tt>FunctionType</tt></a> argument
1355 specifies the formal arguments and return value for the function. The same
1356 <a href="#FunctionTypel"><tt>FunctionType</tt></a> value can be used to
1357 create multiple functions. The <tt>Parent</tt> argument specifies the Module
1358 in which the function is defined. If this argument is provided, the function
1359 will automatically be inserted into that module's list of
1362 <li><tt>bool isExternal()</tt>
1364 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
1365 function is "external", it does not have a body, and thus must be resolved
1366 by linking with a function defined in a different translation unit.</p></li>
1368 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
1369 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
1371 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1372 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1374 <p>These are forwarding methods that make it easy to access the contents of
1375 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
1378 <li><tt>Function::BasicBlockListType &getBasicBlockList()</tt>
1380 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
1381 is necessary to use when you need to update the list or perform a complex
1382 action that doesn't have a forwarding method.</p></li>
1384 <li><tt>Function::aiterator</tt> - Typedef for the argument list
1386 <tt>Function::const_aiterator</tt> - Typedef for const_iterator.<br>
1388 <tt>abegin()</tt>, <tt>aend()</tt>, <tt>afront()</tt>, <tt>aback()</tt>,
1389 <tt>asize()</tt>, <tt>aempty()</tt>, <tt>arbegin()</tt>, <tt>arend()</tt>
1391 <p>These are forwarding methods that make it easy to access the contents of
1392 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
1395 <li><tt>Function::ArgumentListType &getArgumentList()</tt>
1397 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
1398 necessary to use when you need to update the list or perform a complex
1399 action that doesn't have a forwarding method.</p></li>
1401 <li><tt><a href="#BasicBlock">BasicBlock</a> &getEntryBlock()</tt>
1403 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
1404 function. Because the entry block for the function is always the first
1405 block, this returns the first block of the <tt>Function</tt>.</p></li>
1407 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
1408 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
1410 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
1411 <tt>Function</tt> and returns the return type of the function, or the <a
1412 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
1415 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1417 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1418 for this <tt>Function</tt>.</p></li>
1423 <!-- ======================================================================= -->
1424 <div class="doc_subsection">
1425 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
1428 <div class="doc_text">
1431 href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
1433 doxygen info: <a href="/doxygen/classGlobalVariable.html">GlobalVariable
1434 Class</a><br> Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1435 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1437 <p>Global variables are represented with the (suprise suprise)
1438 <tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
1439 subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
1440 always referenced by their address (global values must live in memory, so their
1441 "name" refers to their address). See <a
1442 href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global variables
1443 may have an initial value (which must be a <a
1444 href="#Constant"><tt>Constant</tt></a>), and if they have an initializer, they
1445 may be marked as "constant" themselves (indicating that their contents never
1446 change at runtime).</p>
1450 <!-- _______________________________________________________________________ -->
1451 <div class="doc_subsubsection">
1452 <a name="m_GlobalVariable">Important Public Members of the
1453 <tt>GlobalVariable</tt> class</a>
1456 <div class="doc_text">
1459 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
1460 isConstant, LinkageTypes& Linkage, <a href="#Constant">Constant</a>
1461 *Initializer = 0, const std::string &Name = "", Module* Parent = 0)</tt>
1463 <p>Create a new global variable of the specified type. If
1464 <tt>isConstant</tt> is true then the global variable will be marked as
1465 unchanging for the program. The Linkage parameter specifies the type of
1466 linkage (internal, external, weak, linkonce, appending) for the variable. If
1467 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage, then
1468 the resultant global variable will have internal linkage. AppendingLinkage
1469 concatenates together all instances (in different translation units) of the
1470 variable into a single variable but is only applicable to arrays. See
1471 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
1472 further details on linkage types. Optionally an initializer, a name, and the
1473 module to put the variable into may be specified for the global variable as
1476 <li><tt>bool isConstant() const</tt>
1478 <p>Returns true if this is a global variable that is known not to
1479 be modified at runtime.</p></li>
1481 <li><tt>bool hasInitializer()</tt>
1483 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
1485 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
1487 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
1488 to call this method if there is no initializer.</p></li>
1493 <!-- ======================================================================= -->
1494 <div class="doc_subsection">
1495 <a name="Module">The <tt>Module</tt> class</a>
1498 <div class="doc_text">
1501 href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
1502 <a href="/doxygen/classModule.html">Module Class</a></p>
1504 <p>The <tt>Module</tt> class represents the top level structure present in LLVM
1505 programs. An LLVM module is effectively either a translation unit of the
1506 original program or a combination of several translation units merged by the
1507 linker. The <tt>Module</tt> class keeps track of a list of <a
1508 href="#Function"><tt>Function</tt></a>s, a list of <a
1509 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
1510 href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
1511 helpful member functions that try to make common operations easy.</p>
1515 <!-- _______________________________________________________________________ -->
1516 <div class="doc_subsubsection">
1517 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
1520 <div class="doc_text">
1523 <li><tt>Module::Module(std::string name = "")</tt></li>
1526 <p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
1527 provide a name for it (probably based on the name of the translation unit).</p>
1530 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
1531 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
1533 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1534 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1536 <p>These are forwarding methods that make it easy to access the contents of
1537 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
1540 <li><tt>Module::FunctionListType &getFunctionList()</tt>
1542 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
1543 necessary to use when you need to update the list or perform a complex
1544 action that doesn't have a forwarding method.</p>
1546 <p><!-- Global Variable --></p></li>
1552 <li><tt>Module::giterator</tt> - Typedef for global variable list iterator<br>
1554 <tt>Module::const_giterator</tt> - Typedef for const_iterator.<br>
1556 <tt>gbegin()</tt>, <tt>gend()</tt>, <tt>gfront()</tt>, <tt>gback()</tt>,
1557 <tt>gsize()</tt>, <tt>gempty()</tt>, <tt>grbegin()</tt>, <tt>grend()</tt>
1559 <p> These are forwarding methods that make it easy to access the contents of
1560 a <tt>Module</tt> object's <a
1561 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
1563 <li><tt>Module::GlobalListType &getGlobalList()</tt>
1565 <p>Returns the list of <a
1566 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
1567 use when you need to update the list or perform a complex action that
1568 doesn't have a forwarding method.</p>
1570 <p><!-- Symbol table stuff --> </p></li>
1576 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1578 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1579 for this <tt>Module</tt>.</p>
1581 <p><!-- Convenience methods --></p></li>
1587 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
1588 &Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
1590 <p>Look up the specified function in the <tt>Module</tt> <a
1591 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
1592 <tt>null</tt>.</p></li>
1594 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
1595 std::string &Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
1597 <p>Look up the specified function in the <tt>Module</tt> <a
1598 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
1599 external declaration for the function and return it.</p></li>
1601 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
1603 <p>If there is at least one entry in the <a
1604 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
1605 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
1608 <li><tt>bool addTypeName(const std::string &Name, const <a
1609 href="#Type">Type</a> *Ty)</tt>
1611 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1612 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
1613 name, true is returned and the <a
1614 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
1619 <!-- ======================================================================= -->
1620 <div class="doc_subsection">
1621 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
1624 <div class="doc_text">
1626 <p>Constant represents a base class for different types of constants. It
1627 is subclassed by ConstantBool, ConstantInt, ConstantSInt, ConstantUInt,
1628 ConstantArray etc for representing the various types of Constants.</p>
1632 <!-- _______________________________________________________________________ -->
1633 <div class="doc_subsubsection">
1634 <a name="m_Value">Important Public Methods</a>
1637 <div class="doc_text">
1640 <li><tt>bool isConstantExpr()</tt>: Returns true if it is a
1642 <hr> Important Subclasses of Constant
1645 <li>ConstantSInt : This subclass of Constant represents a signed
1648 <li><tt>int64_t getValue() const</tt>: Returns the underlying value of
1649 this constant. </li>
1652 <li>ConstantUInt : This class represents an unsigned integer.
1654 <li><tt>uint64_t getValue() const</tt>: Returns the underlying value
1655 of this constant. </li>
1658 <li>ConstantFP : This class represents a floating point constant.
1660 <li><tt>double getValue() const</tt>: Returns the underlying value of
1661 this constant. </li>
1664 <li>ConstantBool : This represents a boolean constant.
1666 <li><tt>bool getValue() const</tt>: Returns the underlying value of
1667 this constant. </li>
1670 <li>ConstantArray : This represents a constant array.
1672 <li><tt>const std::vector<Use> &getValues() const</tt>:
1673 Returns a Vecotr of component constants that makeup this array. </li>
1676 <li>ConstantStruct : This represents a constant struct.
1678 <li><tt>const std::vector<Use> &getValues() const</tt>:
1679 Returns a Vecotr of component constants that makeup this array. </li>
1682 <li>ConstantPointerRef : This represents a constant pointer value
1683 that is initialized to point to a global value, which lies at a
1684 constant fixed address.
1686 <li><tt>GlobalValue *getValue()</tt>: Returns the global
1687 value to which this pointer is pointing to. </li>
1696 <!-- ======================================================================= -->
1697 <div class="doc_subsection">
1698 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
1701 <div class="doc_text">
1703 <p>Type as noted earlier is also a subclass of a Value class. Any primitive
1704 type (like int, short etc) in LLVM is an instance of Type Class. All other
1705 types are instances of subclasses of type like FunctionType, ArrayType
1706 etc. DerivedType is the interface for all such dervied types including
1707 FunctionType, ArrayType, PointerType, StructType. Types can have names. They can
1708 be recursive (StructType). There exists exactly one instance of any type
1709 structure at a time. This allows using pointer equality of Type *s for comparing
1714 <!-- _______________________________________________________________________ -->
1715 <div class="doc_subsubsection">
1716 <a name="m_Value">Important Public Methods</a>
1719 <div class="doc_text">
1723 <li><tt>bool isSigned() const</tt>: Returns whether an integral numeric type
1724 is signed. This is true for SByteTy, ShortTy, IntTy, LongTy. Note that this is
1725 not true for Float and Double. </li>
1727 <li><tt>bool isUnsigned() const</tt>: Returns whether a numeric type is
1728 unsigned. This is not quite the complement of isSigned... nonnumeric types
1729 return false as they do with isSigned. This returns true for UByteTy,
1730 UShortTy, UIntTy, and ULongTy. </li>
1732 <li><tt>bool isInteger() const</tt>: Equilivent to isSigned() || isUnsigned(),
1733 but with only a single virtual function invocation.</li>
1735 <li><tt>bool isIntegral() const</tt>: Returns true if this is an integral
1736 type, which is either Bool type or one of the Integer types.</li>
1738 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
1739 floating point types.</li>
1741 <li><tt>isLosslesslyConvertableTo (const Type *Ty) const</tt>: Return true if
1742 this type can be converted to 'Ty' without any reinterpretation of bits. For
1743 example, uint to int or one pointer type to another.</li>
1746 <p>Derived Types</p>
1749 <li>SequentialType : This is subclassed by ArrayType and PointerType
1751 <li><tt>const Type * getElementType() const</tt>: Returns the type of
1752 each of the elements in the sequential type. </li>
1755 <li>ArrayType : This is a subclass of SequentialType and defines
1756 interface for array types.
1758 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
1759 elements in the array. </li>
1762 <li>PointerType : Subclass of SequentialType for pointer types. </li>
1763 <li>StructType : subclass of DerivedTypes for struct types </li>
1764 <li>FunctionType : subclass of DerivedTypes for function types.
1766 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
1768 <li><tt> const Type * getReturnType() const</tt>: Returns the
1769 return type of the function.</li>
1770 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
1771 the type of the ith parameter.</li>
1772 <li><tt> const unsigned getNumParams() const</tt>: Returns the
1773 number of formal parameters.</li>
1782 <!-- ======================================================================= -->
1783 <div class="doc_subsection">
1784 <a name="Argument">The <tt>Argument</tt> class</a>
1787 <div class="doc_text">
1789 <p>This subclass of Value defines the interface for incoming formal
1790 arguments to a function. A Function maitanis a list of its formal
1791 arguments. An argument has a pointer to the parent Function.</p>
1795 <!-- ======================================================================= -->
1796 <div class="doc_subsection">
1797 <a name="SymbolTable">The <tt>SymbolTable</tt> class</a>
1799 <div class="doc_text">
1800 <p>This class provides a symbol table that the
1801 <a href="#Function"><tt>Function</tt></a> and <a href="#Module">
1802 <tt>Module</tt></a> classes use for naming definitions. The symbol table can
1803 provide a name for any <a href="#Value"><tt>Value</tt></a> or
1804 <a href="#Type"><tt>Type</tt></a>. <tt>SymbolTable</tt> is an abstract data
1805 type. It hides the data it contains and provides access to it through a
1806 controlled interface.</p>
1808 <p>To use the <tt>SymbolTable</tt> well, you need to understand the
1809 structure of the information it holds. The class contains two
1810 <tt>std::map</tt> objects. The first, <tt>pmap</tt>, is a map of
1811 <tt>Type*</tt> to maps of name (<tt>std::string</tt>) to <tt>Value*</tt>.
1812 The second, <tt>tmap</tt>, is a map of names to <tt>Type*</tt>. Thus, Values
1813 are stored in two-dimensions and accessed by <tt>Type</tt> and name. Types,
1814 however, are stored in a single dimension and accessed only by name.</p>
1816 <p>The interface of this class provides three basic types of operations:
1818 <li><em>Accessors</em>. Accessors provide read-only access to information
1819 such as finding a value for a name with the
1820 <a href="#SymbolTable_lookup">lookup</a> method.</li>
1821 <li><em>Mutators</em>. Mutators allow the user to add information to the
1822 <tt>SymbolTable</tt> with methods like
1823 <a href="#SymbolTable_insert"><tt>insert</tt></a>.</li>
1824 <li><em>Iterators</em>. Iterators allow the user to traverse the content
1825 of the symbol table in well defined ways, such as the method
1826 <a href="#SymbolTable_type_begin"><tt>type_begin</tt></a>.</li>
1831 <dt><tt>Value* lookup(const Type* Ty, const std::string& name) const</tt>:
1833 <dd>The <tt>lookup</tt> method searches the type plane given by the
1834 <tt>Ty</tt> parameter for a <tt>Value</tt> with the provided <tt>name</tt>.
1835 If a suitable <tt>Value</tt> is not found, null is returned.</dd>
1837 <dt><tt>Type* lookupType( const std::string& name) const</tt>:</dt>
1838 <dd>The <tt>lookupType</tt> method searches through the types for a
1839 <tt>Type</tt> with the provided <tt>name</tt>. If a suitable <tt>Type</tt>
1840 is not found, null is returned.</dd>
1842 <dt><tt>bool hasTypes() const</tt>:</dt>
1843 <dd>This function returns true if an entry has been made into the type
1846 <dt><tt>bool isEmpty() const</tt>:</dt>
1847 <dd>This function returns true if both the value and types maps are
1850 <dt><tt>std::string get_name(const Value*) const</tt>:</dt>
1851 <dd>This function returns the name of the Value provided or the empty
1852 string if the Value is not in the symbol table.</dd>
1854 <dt><tt>std::string get_name(const Type*) const</tt>:</dt>
1855 <dd>This function returns the name of the Type provided or the empty
1856 string if the Type is not in the symbol table.</dd>
1861 <dt><tt>void insert(Value *Val)</tt>:</dt>
1862 <dd>This method adds the provided value to the symbol table. The Value must
1863 have both a name and a type which are extracted and used to place the value
1864 in the correct type plane under the value's name.</dd>
1866 <dt><tt>void insert(const std::string& Name, Value *Val)</tt>:</dt>
1867 <dd> Inserts a constant or type into the symbol table with the specified
1868 name. There can be a many to one mapping between names and constants
1871 <dt><tt>void insert(const std::string& Name, Type *Typ)</tt>:</dt>
1872 <dd> Inserts a type into the symbol table with the specified name. There
1873 can be a many-to-one mapping between names and types. This method
1874 allows a type with an existing entry in the symbol table to get
1877 <dt><tt>void remove(Value* Val)</tt>:</dt>
1878 <dd> This method removes a named value from the symbol table. The
1879 type and name of the Value are extracted from \p N and used to
1880 lookup the Value in the correct type plane. If the Value is
1881 not in the symbol table, this method silently ignores the
1884 <dt><tt>void remove(Type* Typ)</tt>:</dt>
1885 <dd> This method removes a named type from the symbol table. The
1886 name of the type is extracted from \P T and used to look up
1887 the Type in the type map. If the Type is not in the symbol
1888 table, this method silently ignores the request.</dd>
1890 <dt><tt>Value* remove(const std::string& Name, Value *Val)</tt>:</dt>
1891 <dd> Remove a constant or type with the specified name from the
1894 <dt><tt>Type* remove(const std::string& Name, Type* T)</tt>:</dt>
1895 <dd> Remove a type with the specified name from the symbol table.
1896 Returns the removed Type.</dd>
1898 <dt><tt>Value *value_remove(const value_iterator& It)</tt>:</dt>
1899 <dd> Removes a specific value from the symbol table.
1900 Returns the removed value.</dd>
1902 <dt><tt>bool strip()</tt>:</dt>
1903 <dd> This method will strip the symbol table of its names leaving
1904 the type and values. </dd>
1906 <dt><tt>void clear()</tt>:</dt>
1907 <dd>Empty the symbol table completely.</dd>
1911 <p>The following functions describe three types of iterators you can obtain
1912 the beginning or end of the sequence for both const and non-const. It is
1913 important to keep track of the different kinds of iterators. There are
1914 three idioms worth pointing out:</p>
1915 <table class="doc_table">
1916 <tr><th>Units</th><th>Iterator</th><th>Idiom</th></tr>
1918 <td>Planes Of name/Value maps</td><td>PI</td>
1920 for (SymbolTable::plane_const_iterator PI = ST.plane_begin(),
1921 PE = ST.plane_end(); PI != PE; ++PI ) {
1922 PI->first // This is the Type* of the plane
1923 PI->second // This is the SymbolTable::ValueMap of name/Value pairs
1927 <td>All name/Type Pairs</td><td>TI</td>
1929 for (SymbolTable::type_const_iterator TI = ST.type_begin(),
1930 TE = ST.type_end(); TI != TE; ++TI )
1931 TI->first // This is the name of the type
1932 TI->second // This is the Type* value associated with the name
1936 <td>name/Value pairs in a plane</td><td>VI</td>
1938 for (SymbolTable::value_const_iterator VI = ST.value_begin(SomeType),
1939 VE = ST.value_end(SomeType); VI != VE; ++VI )
1940 VI->first // This is the name of the Value
1941 VI->second // This is the Value* value associated with the name
1945 <p>Using the recommended iterator names and idioms will help you avoid
1946 making mistakes. Of particular note, make sure that whenever you use
1947 value_begin(SomeType) that you always compare the resulting iterator
1948 with value_end(SomeType) not value_end(SomeOtherType) or else you
1949 will loop infinitely.</p>
1953 <dt><tt>plane_iterator plane_begin()</tt>:</dt>
1954 <dd>Get an iterator that starts at the beginning of the type planes.
1955 The iterator will iterate over the Type/ValueMap pairs in the
1958 <dt><tt>plane_const_iterator plane_begin() const</tt>:</dt>
1959 <dd>Get a const_iterator that starts at the beginning of the type
1960 planes. The iterator will iterate over the Type/ValueMap pairs
1961 in the type planes. </dd>
1963 <dt><tt>plane_iterator plane_end()</tt>:</dt>
1964 <dd>Get an iterator at the end of the type planes. This serves as
1965 the marker for end of iteration over the type planes.</dd>
1967 <dt><tt>plane_const_iterator plane_end() const</tt>:</dt>
1968 <dd>Get a const_iterator at the end of the type planes. This serves as
1969 the marker for end of iteration over the type planes.</dd>
1971 <dt><tt>value_iterator value_begin(const Type *Typ)</tt>:</dt>
1972 <dd>Get an iterator that starts at the beginning of a type plane.
1973 The iterator will iterate over the name/value pairs in the type plane.
1974 Note: The type plane must already exist before using this.</dd>
1976 <dt><tt>value_const_iterator value_begin(const Type *Typ) const</tt>:</dt>
1977 <dd>Get a const_iterator that starts at the beginning of a type plane.
1978 The iterator will iterate over the name/value pairs in the type plane.
1979 Note: The type plane must already exist before using this.</dd>
1981 <dt><tt>value_iterator value_end(const Type *Typ)</tt>:</dt>
1982 <dd>Get an iterator to the end of a type plane. This serves as the marker
1983 for end of iteration of the type plane.
1984 Note: The type plane must already exist before using this.</dd>
1986 <dt><tt>value_const_iterator value_end(const Type *Typ) const</tt>:</dt>
1987 <dd>Get a const_iterator to the end of a type plane. This serves as the
1988 marker for end of iteration of the type plane.
1989 Note: the type plane must already exist before using this.</dd>
1991 <dt><tt>type_iterator type_begin()</tt>:</dt>
1992 <dd>Get an iterator to the start of the name/Type map.</dd>
1994 <dt><tt>type_const_iterator type_begin() cons</tt>:</dt>
1995 <dd> Get a const_iterator to the start of the name/Type map.</dd>
1997 <dt><tt>type_iterator type_end()</tt>:</dt>
1998 <dd>Get an iterator to the end of the name/Type map. This serves as the
1999 marker for end of iteration of the types.</dd>
2001 <dt><tt>type_const_iterator type_end() const</tt>:</dt>
2002 <dd>Get a const-iterator to the end of the name/Type map. This serves
2003 as the marker for end of iteration of the types.</dd>
2005 <dt><tt>plane_const_iterator find(const Type* Typ ) const</tt>:</dt>
2006 <dd>This method returns a plane_const_iterator for iteration over
2007 the type planes starting at a specific plane, given by \p Ty.</dd>
2009 <dt><tt>plane_iterator find( const Type* Typ </tt>:</dt>
2010 <dd>This method returns a plane_iterator for iteration over the
2011 type planes starting at a specific plane, given by \p Ty.</dd>
2013 <dt><tt>const ValueMap* findPlane( const Type* Typ ) cons</tt>:</dt>
2014 <dd>This method returns a ValueMap* for a specific type plane. This
2015 interface is deprecated and may go away in the future.</dd>
2019 <!-- *********************************************************************** -->
2022 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
2023 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
2024 <a href="http://validator.w3.org/check/referer"><img
2025 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
2027 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
2028 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
2029 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
2030 Last modified: $Date$