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="#Module">The <tt>Module</tt> class</a></li>
93 <li><a href="#Constant">The <tt>Constant</tt> class</a>
95 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
97 <li><a href="#BasicBlock">The <tt>BasicBlock</tt>class</a></li>
98 <li><a href="#Function">The <tt>Function</tt> class</a></li>
99 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class
104 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
105 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
108 <li><a href="#SymbolTable">The <tt>SymbolTable</tt> class </a></li>
109 <li>The <tt>ilist</tt> and <tt>iplist</tt> classes
111 <li>Creating, inserting, moving and deleting from LLVM lists </li>
114 <li>Important iterator invalidation semantics to be aware of.</li>
118 <div class="doc_author">
119 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
120 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
121 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>, and
122 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a></p>
125 <!-- *********************************************************************** -->
126 <div class="doc_section">
127 <a name="introduction">Introduction </a>
129 <!-- *********************************************************************** -->
131 <div class="doc_text">
133 <p>This document is meant to highlight some of the important classes and
134 interfaces available in the LLVM source-base. This manual is not
135 intended to explain what LLVM is, how it works, and what LLVM code looks
136 like. It assumes that you know the basics of LLVM and are interested
137 in writing transformations or otherwise analyzing or manipulating the
140 <p>This document should get you oriented so that you can find your
141 way in the continuously growing source code that makes up the LLVM
142 infrastructure. Note that this manual is not intended to serve as a
143 replacement for reading the source code, so if you think there should be
144 a method in one of these classes to do something, but it's not listed,
145 check the source. Links to the <a href="/doxygen/">doxygen</a> sources
146 are provided to make this as easy as possible.</p>
148 <p>The first section of this document describes general information that is
149 useful to know when working in the LLVM infrastructure, and the second describes
150 the Core LLVM classes. In the future this manual will be extended with
151 information describing how to use extension libraries, such as dominator
152 information, CFG traversal routines, and useful utilities like the <tt><a
153 href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
157 <!-- *********************************************************************** -->
158 <div class="doc_section">
159 <a name="general">General Information</a>
161 <!-- *********************************************************************** -->
163 <div class="doc_text">
165 <p>This section contains general information that is useful if you are working
166 in the LLVM source-base, but that isn't specific to any particular API.</p>
170 <!-- ======================================================================= -->
171 <div class="doc_subsection">
172 <a name="stl">The C++ Standard Template Library</a>
175 <div class="doc_text">
177 <p>LLVM makes heavy use of the C++ Standard Template Library (STL),
178 perhaps much more than you are used to, or have seen before. Because of
179 this, you might want to do a little background reading in the
180 techniques used and capabilities of the library. There are many good
181 pages that discuss the STL, and several books on the subject that you
182 can get, so it will not be discussed in this document.</p>
184 <p>Here are some useful links:</p>
188 <li><a href="http://www.dinkumware.com/refxcpp.html">Dinkumware C++ Library
189 reference</a> - an excellent reference for the STL and other parts of the
190 standard C++ library.</li>
192 <li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
193 O'Reilly book in the making. It has a decent
195 Reference that rivals Dinkumware's, and is unfortunately no longer free since the book has been
198 <li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
201 <li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
203 href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
206 <li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
209 <li><a href="http://www.linux.com.cn/Bruce_Eckel/TICPPv2/Contents.htm">
210 Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
215 <p>You are also encouraged to take a look at the <a
216 href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
217 to write maintainable code more than where to put your curly braces.</p>
221 <!-- ======================================================================= -->
222 <div class="doc_subsection">
223 <a name="stl">Other useful references</a>
226 <div class="doc_text">
229 <li><a href="http://www.psc.edu/%7Esemke/cvs_branches.html">CVS
230 Branch and Tag Primer</a></li>
231 <li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
232 static and shared libraries across platforms</a></li>
237 <!-- *********************************************************************** -->
238 <div class="doc_section">
239 <a name="apis">Important and useful LLVM APIs</a>
241 <!-- *********************************************************************** -->
243 <div class="doc_text">
245 <p>Here we highlight some LLVM APIs that are generally useful and good to
246 know about when writing transformations.</p>
250 <!-- ======================================================================= -->
251 <div class="doc_subsection">
252 <a name="isa">The isa<>, cast<> and dyn_cast<> templates</a>
255 <div class="doc_text">
257 <p>The LLVM source-base makes extensive use of a custom form of RTTI.
258 These templates have many similarities to the C++ <tt>dynamic_cast<></tt>
259 operator, but they don't have some drawbacks (primarily stemming from
260 the fact that <tt>dynamic_cast<></tt> only works on classes that
261 have a v-table). Because they are used so often, you must know what they
262 do and how they work. All of these templates are defined in the <a
263 href="/doxygen/Casting_8h-source.html"><tt>Support/Casting.h</tt></a>
264 file (note that you very rarely have to include this file directly).</p>
267 <dt><tt>isa<></tt>: </dt>
269 <dd>The <tt>isa<></tt> operator works exactly like the Java
270 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
271 a reference or pointer points to an instance of the specified class. This can
272 be very useful for constraint checking of various sorts (example below).</dd>
274 <dt><tt>cast<></tt>: </dt>
276 <dd>The <tt>cast<></tt> operator is a "checked cast" operation. It
277 converts a pointer or reference from a base class to a derived cast, causing
278 an assertion failure if it is not really an instance of the right type. This
279 should be used in cases where you have some information that makes you believe
280 that something is of the right type. An example of the <tt>isa<></tt>
281 and <tt>cast<></tt> template is:
284 static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
285 if (isa<<a href="#Constant">Constant</a>>(V) || isa<<a href="#Argument">Argument</a>>(V) || isa<<a href="#GlobalValue">GlobalValue</a>>(V))
288 <i>// Otherwise, it must be an instruction...</i>
289 return !L->contains(cast<<a href="#Instruction">Instruction</a>>(V)->getParent());
292 <p>Note that you should <b>not</b> use an <tt>isa<></tt> test followed
293 by a <tt>cast<></tt>, for that use the <tt>dyn_cast<></tt>
298 <dt><tt>dyn_cast<></tt>:</dt>
300 <dd>The <tt>dyn_cast<></tt> operator is a "checking cast" operation. It
301 checks to see if the operand is of the specified type, and if so, returns a
302 pointer to it (this operator does not work with references). If the operand is
303 not of the correct type, a null pointer is returned. Thus, this works very
304 much like the <tt>dynamic_cast</tt> operator in C++, and should be used in the
305 same circumstances. Typically, the <tt>dyn_cast<></tt> operator is used
306 in an <tt>if</tt> statement or some other flow control statement like this:
309 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast<<a href="#AllocationInst">AllocationInst</a>>(Val)) {
314 <p> This form of the <tt>if</tt> statement effectively combines together a
315 call to <tt>isa<></tt> and a call to <tt>cast<></tt> into one
316 statement, which is very convenient.</p>
318 <p> Another common example is:</p>
321 <i>// Loop over all of the phi nodes in a basic block</i>
322 BasicBlock::iterator BBI = BB->begin();
323 for (; <a href="#PhiNode">PHINode</a> *PN = dyn_cast<<a href="#PHINode">PHINode</a>>(BBI); ++BBI)
324 std::cerr << *PN;
327 <p>Note that the <tt>dyn_cast<></tt> operator, like C++'s
328 <tt>dynamic_cast</tt> or Java's <tt>instanceof</tt> operator, can be abused.
329 In particular you should not use big chained <tt>if/then/else</tt> blocks to
330 check for lots of different variants of classes. If you find yourself
331 wanting to do this, it is much cleaner and more efficient to use the
332 InstVisitor class to dispatch over the instruction type directly.</p>
336 <dt><tt>cast_or_null<></tt>: </dt>
338 <dd>The <tt>cast_or_null<></tt> operator works just like the
339 <tt>cast<></tt> operator, except that it allows for a null pointer as
340 an argument (which it then propagates). This can sometimes be useful,
341 allowing you to combine several null checks into one.</dd>
343 <dt><tt>dyn_cast_or_null<></tt>: </dt>
345 <dd>The <tt>dyn_cast_or_null<></tt> operator works just like the
346 <tt>dyn_cast<></tt> operator, except that it allows for a null pointer
347 as an argument (which it then propagates). This can sometimes be useful,
348 allowing you to combine several null checks into one.</dd>
352 <p>These five templates can be used with any classes, whether they have a
353 v-table or not. To add support for these templates, you simply need to add
354 <tt>classof</tt> static methods to the class you are interested casting
355 to. Describing this is currently outside the scope of this document, but there
356 are lots of examples in the LLVM source base.</p>
360 <!-- ======================================================================= -->
361 <div class="doc_subsection">
362 <a name="DEBUG">The <tt>DEBUG()</tt> macro & <tt>-debug</tt> option</a>
365 <div class="doc_text">
367 <p>Often when working on your pass you will put a bunch of debugging printouts
368 and other code into your pass. After you get it working, you want to remove
369 it... but you may need it again in the future (to work out new bugs that you run
372 <p> Naturally, because of this, you don't want to delete the debug printouts,
373 but you don't want them to always be noisy. A standard compromise is to comment
374 them out, allowing you to enable them if you need them in the future.</p>
376 <p>The "<tt><a href="/doxygen/Debug_8h-source.html">Support/Debug.h</a></tt>"
377 file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
378 this problem. Basically, you can put arbitrary code into the argument of the
379 <tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
380 tool) is run with the '<tt>-debug</tt>' command line argument:</p>
382 <pre> ... <br> DEBUG(std::cerr << "I am here!\n");<br> ...<br></pre>
384 <p>Then you can run your pass like this:</p>
386 <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>
388 <p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
389 to not have to create "yet another" command line option for the debug output for
390 your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
391 so they do not cause a performance impact at all (for the same reason, they
392 should also not contain side-effects!).</p>
394 <p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
395 enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
396 "<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
397 program hasn't been started yet, you can always just run it with
402 <!-- _______________________________________________________________________ -->
403 <div class="doc_subsubsection">
404 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE()</tt> and
405 the <tt>-debug-only</tt> option</a>
408 <div class="doc_text">
410 <p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
411 just turns on <b>too much</b> information (such as when working on the code
412 generator). If you want to enable debug information with more fine-grained
413 control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
414 option as follows:</p>
416 <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>
418 <p>Then you can run your pass like this:</p>
420 <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>
422 <p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
423 a file, to specify the debug type for the entire module (if you do this before
424 you <tt>#include "Support/Debug.h"</tt>, you don't have to insert the ugly
425 <tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
426 "bar", because there is no system in place to ensure that names do not
427 conflict. If two different modules use the same string, they will all be turned
428 on when the name is specified. This allows, for example, all debug information
429 for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
430 even if the source lives in multiple files.</p>
434 <!-- ======================================================================= -->
435 <div class="doc_subsection">
436 <a name="Statistic">The <tt>Statistic</tt> template & <tt>-stats</tt>
440 <div class="doc_text">
443 href="/doxygen/Statistic_8h-source.html">Support/Statistic.h</a></tt>" file
444 provides a template named <tt>Statistic</tt> that is used as a unified way to
445 keep track of what the LLVM compiler is doing and how effective various
446 optimizations are. It is useful to see what optimizations are contributing to
447 making a particular program run faster.</p>
449 <p>Often you may run your pass on some big program, and you're interested to see
450 how many times it makes a certain transformation. Although you can do this with
451 hand inspection, or some ad-hoc method, this is a real pain and not very useful
452 for big programs. Using the <tt>Statistic</tt> template makes it very easy to
453 keep track of this information, and the calculated information is presented in a
454 uniform manner with the rest of the passes being executed.</p>
456 <p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
457 it are as follows:</p>
460 <li>Define your statistic like this:
461 <pre>static Statistic<> NumXForms("mypassname", "The # of times I did stuff");<br></pre>
463 <p>The <tt>Statistic</tt> template can emulate just about any data-type,
464 but if you do not specify a template argument, it defaults to acting like
465 an unsigned int counter (this is usually what you want).</p></li>
467 <li>Whenever you make a transformation, bump the counter:
468 <pre> ++NumXForms; // I did stuff<br></pre>
472 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
473 statistics gathered, use the '<tt>-stats</tt>' option:</p>
475 <pre> $ opt -stats -mypassname < program.bc > /dev/null<br> ... statistic output ...<br></pre>
477 <p> When running <tt>gccas</tt> on a C file from the SPEC benchmark
478 suite, it gives a report that looks like this:</p>
480 <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>
482 <p>Obviously, with so many optimizations, having a unified framework for this
483 stuff is very nice. Making your pass fit well into the framework makes it more
484 maintainable and useful.</p>
488 <!-- *********************************************************************** -->
489 <div class="doc_section">
490 <a name="common">Helpful Hints for Common Operations</a>
492 <!-- *********************************************************************** -->
494 <div class="doc_text">
496 <p>This section describes how to perform some very simple transformations of
497 LLVM code. This is meant to give examples of common idioms used, showing the
498 practical side of LLVM transformations. <p> Because this is a "how-to" section,
499 you should also read about the main classes that you will be working with. The
500 <a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
501 and descriptions of the main classes that you should know about.</p>
505 <!-- NOTE: this section should be heavy on example code -->
506 <!-- ======================================================================= -->
507 <div class="doc_subsection">
508 <a name="inspection">Basic Inspection and Traversal Routines</a>
511 <div class="doc_text">
513 <p>The LLVM compiler infrastructure have many different data structures that may
514 be traversed. Following the example of the C++ standard template library, the
515 techniques used to traverse these various data structures are all basically the
516 same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
517 method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
518 function returns an iterator pointing to one past the last valid element of the
519 sequence, and there is some <tt>XXXiterator</tt> data type that is common
520 between the two operations.</p>
522 <p>Because the pattern for iteration is common across many different aspects of
523 the program representation, the standard template library algorithms may be used
524 on them, and it is easier to remember how to iterate. First we show a few common
525 examples of the data structures that need to be traversed. Other data
526 structures are traversed in very similar ways.</p>
530 <!-- _______________________________________________________________________ -->
531 <div class="doc_subsubsection">
532 <a name="iterate_function">Iterating over the </a><a
533 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
534 href="#Function"><tt>Function</tt></a>
537 <div class="doc_text">
539 <p>It's quite common to have a <tt>Function</tt> instance that you'd like to
540 transform in some way; in particular, you'd like to manipulate its
541 <tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
542 the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
543 an example that prints the name of a <tt>BasicBlock</tt> and the number of
544 <tt>Instruction</tt>s it contains:</p>
546 <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>
548 <p>Note that i can be used as if it were a pointer for the purposes of
549 invoking member functions of the <tt>Instruction</tt> class. This is
550 because the indirection operator is overloaded for the iterator
551 classes. In the above code, the expression <tt>i->size()</tt> is
552 exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
556 <!-- _______________________________________________________________________ -->
557 <div class="doc_subsubsection">
558 <a name="iterate_basicblock">Iterating over the </a><a
559 href="#Instruction"><tt>Instruction</tt></a>s in a <a
560 href="#BasicBlock"><tt>BasicBlock</tt></a>
563 <div class="doc_text">
565 <p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
566 easy to iterate over the individual instructions that make up
567 <tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
568 a <tt>BasicBlock</tt>:</p>
570 <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>
572 <p>However, this isn't really the best way to print out the contents of a
573 <tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
574 anything you'll care about, you could have just invoked the print routine on the
575 basic block itself: <tt>cerr << *blk << "\n";</tt>.</p>
577 <p>Note that currently operator<< is implemented for <tt>Value*</tt>, so
578 it will print out the contents of the pointer, instead of the pointer value you
579 might expect. This is a deprecated interface that will be removed in the
580 future, so it's best not to depend on it. To print out the pointer value for
581 now, you must cast to <tt>void*</tt>.</p>
585 <!-- _______________________________________________________________________ -->
586 <div class="doc_subsubsection">
587 <a name="iterate_institer">Iterating over the </a><a
588 href="#Instruction"><tt>Instruction</tt></a>s in a <a
589 href="#Function"><tt>Function</tt></a>
592 <div class="doc_text">
594 <p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
595 <tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
596 <tt>InstIterator</tt> should be used instead. You'll need to include <a
597 href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
598 and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
599 small example that shows how to dump all instructions in a function to the standard error stream:<p>
601 <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>
602 Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
603 worklist with its initial contents. For example, if you wanted to
604 initialize a worklist to contain all instructions in a <tt>Function</tt>
605 F, all you would need to do is something like:
606 <pre>std::set<Instruction*> worklist;<br>worklist.insert(inst_begin(F), inst_end(F));<br></pre>
608 <p>The STL set <tt>worklist</tt> would now contain all instructions in the
609 <tt>Function</tt> pointed to by F.</p>
613 <!-- _______________________________________________________________________ -->
614 <div class="doc_subsubsection">
615 <a name="iterate_convert">Turning an iterator into a class pointer (and
619 <div class="doc_text">
621 <p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
622 instance when all you've got at hand is an iterator. Well, extracting
623 a reference or a pointer from an iterator is very straight-forward.
624 Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
625 is a <tt>BasicBlock::const_iterator</tt>:</p>
627 <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>
629 <p>However, the iterators you'll be working with in the LLVM framework are
630 special: they will automatically convert to a ptr-to-instance type whenever they
631 need to. Instead of dereferencing the iterator and then taking the address of
632 the result, you can simply assign the iterator to the proper pointer type and
633 you get the dereference and address-of operation as a result of the assignment
634 (behind the scenes, this is a result of overloading casting mechanisms). Thus
635 the last line of the last example,</p>
637 <pre>Instruction* pinst = &*i;</pre>
639 <p>is semantically equivalent to</p>
641 <pre>Instruction* pinst = i;</pre>
643 <p>It's also possible to turn a class pointer into the corresponding iterator,
644 and this is a constant time operation (very efficient). The following code
645 snippet illustrates use of the conversion constructors provided by LLVM
646 iterators. By using these, you can explicitly grab the iterator of something
647 without actually obtaining it via iteration over some structure:</p>
649 <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>
653 <!--_______________________________________________________________________-->
654 <div class="doc_subsubsection">
655 <a name="iterate_complex">Finding call sites: a slightly more complex
659 <div class="doc_text">
661 <p>Say that you're writing a FunctionPass and would like to count all the
662 locations in the entire module (that is, across every <tt>Function</tt>) where a
663 certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
664 learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
665 much more straight-forward manner, but this example will allow us to explore how
666 you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudocode, this
667 is what we want to do:</p>
669 <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>
671 <p>And the actual code is (remember, since we're writing a
672 <tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
673 override the <tt>runOnFunction</tt> method...):</p>
675 <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
676 href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a><<a
677 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>
681 <!--_______________________________________________________________________-->
682 <div class="doc_subsubsection">
683 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
686 <div class="doc_text">
688 <p>You may have noticed that the previous example was a bit oversimplified in
689 that it did not deal with call sites generated by 'invoke' instructions. In
690 this, and in other situations, you may find that you want to treat
691 <tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
692 most-specific common base class is <tt>Instruction</tt>, which includes lots of
693 less closely-related things. For these cases, LLVM provides a handy wrapper
695 href="http://llvm.cs.uiuc.edu/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
696 It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
697 methods that provide functionality common to <tt>CallInst</tt>s and
698 <tt>InvokeInst</tt>s.</p>
700 <p>This class has "value semantics": it should be passed by value, not by
701 reference and it should not be dynamically allocated or deallocated using
702 <tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
703 assignable and constructable, with costs equivalents to that of a bare pointer.
704 If you look at its definition, it has only a single pointer member.</p>
708 <!--_______________________________________________________________________-->
709 <div class="doc_subsubsection">
710 <a name="iterate_chains">Iterating over def-use & use-def chains</a>
713 <div class="doc_text">
715 <p>Frequently, we might have an instance of the <a
716 href="/doxygen/structllvm_1_1Value.html">Value Class</a> and we want to
717 determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
718 <tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
719 For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
720 particular function <tt>foo</tt>. Finding all of the instructions that
721 <i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
724 <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>
726 <p>Alternately, it's common to have an instance of the <a
727 href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
728 <tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
729 <tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
730 <tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
731 all of the values that a particular instruction uses (that is, the operands of
732 the particular <tt>Instruction</tt>):</p>
734 <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>
737 def-use chains ("finding all users of"): Value::use_begin/use_end
738 use-def chains ("finding all values used"): User::op_begin/op_end [op=operand]
743 <!-- ======================================================================= -->
744 <div class="doc_subsection">
745 <a name="simplechanges">Making simple changes</a>
748 <div class="doc_text">
750 <p>There are some primitive transformation operations present in the LLVM
751 infrastructure that are worth knowing about. When performing
752 transformations, it's fairly common to manipulate the contents of basic
753 blocks. This section describes some of the common methods for doing so
754 and gives example code.</p>
758 <!--_______________________________________________________________________-->
759 <div class="doc_subsubsection">
760 <a name="schanges_creating">Creating and inserting new
761 <tt>Instruction</tt>s</a>
764 <div class="doc_text">
766 <p><i>Instantiating Instructions</i></p>
768 <p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
769 constructor for the kind of instruction to instantiate and provide the necessary
770 parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
771 (const-ptr-to) <tt>Type</tt>. Thus:</p>
773 <pre>AllocaInst* ai = new AllocaInst(Type::IntTy);</pre>
775 <p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
776 one integer in the current stack frame, at runtime. Each <tt>Instruction</tt>
777 subclass is likely to have varying default parameters which change the semantics
778 of the instruction, so refer to the <a
779 href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
780 Instruction</a> that you're interested in instantiating.</p>
782 <p><i>Naming values</i></p>
784 <p>It is very useful to name the values of instructions when you're able to, as
785 this facilitates the debugging of your transformations. If you end up looking
786 at generated LLVM machine code, you definitely want to have logical names
787 associated with the results of instructions! By supplying a value for the
788 <tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
789 associate a logical name with the result of the instruction's execution at
790 runtime. For example, say that I'm writing a transformation that dynamically
791 allocates space for an integer on the stack, and that integer is going to be
792 used as some kind of index by some other code. To accomplish this, I place an
793 <tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
794 <tt>Function</tt>, and I'm intending to use it within the same
795 <tt>Function</tt>. I might do:</p>
797 <pre>AllocaInst* pa = new AllocaInst(Type::IntTy, 0, "indexLoc");</pre>
799 <p>where <tt>indexLoc</tt> is now the logical name of the instruction's
800 execution value, which is a pointer to an integer on the runtime stack.</p>
802 <p><i>Inserting instructions</i></p>
804 <p>There are essentially two ways to insert an <tt>Instruction</tt>
805 into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
808 <li>Insertion into an explicit instruction list
810 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
811 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
812 before <tt>*pi</tt>, we do the following: </p>
814 <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>
816 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
817 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
818 classes provide constructors which take a pointer to a
819 <tt>BasicBlock</tt> to be appended to. For example code that
822 <pre> BasicBlock *pb = ...;<br> Instruction *newInst = new Instruction(...);<br> pb->getInstList().push_back(newInst); // appends newInst to pb<br></pre>
826 <pre> BasicBlock *pb = ...;<br> Instruction *newInst = new Instruction(..., pb);<br></pre>
828 <p>which is much cleaner, especially if you are creating
829 long instruction streams.</p></li>
831 <li>Insertion into an implicit instruction list
833 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
834 are implicitly associated with an existing instruction list: the instruction
835 list of the enclosing basic block. Thus, we could have accomplished the same
836 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
839 <pre> Instruction *pi = ...;<br> Instruction *newInst = new Instruction(...);<br> pi->getParent()->getInstList().insert(pi, newInst);<br></pre>
841 <p>In fact, this sequence of steps occurs so frequently that the
842 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
843 constructors which take (as a default parameter) a pointer to an
844 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
845 precede. That is, <tt>Instruction</tt> constructors are capable of
846 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
847 provided instruction, immediately before that instruction. Using an
848 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
849 parameter, the above code becomes:</p>
851 <pre>Instruction* pi = ...;<br>Instruction* newInst = new Instruction(..., pi);<br></pre>
853 <p>which is much cleaner, especially if you're creating a lot of
854 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
859 <!--_______________________________________________________________________-->
860 <div class="doc_subsubsection">
861 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
864 <div class="doc_text">
866 <p>Deleting an instruction from an existing sequence of instructions that form a
867 <a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward. First,
868 you must have a pointer to the instruction that you wish to delete. Second, you
869 need to obtain the pointer to that instruction's basic block. You use the
870 pointer to the basic block to get its list of instructions and then use the
871 erase function to remove your instruction. For example:</p>
873 <pre> <a href="#Instruction">Instruction</a> *I = .. ;<br> <a
874 href="#BasicBlock">BasicBlock</a> *BB = I->getParent();<br> BB->getInstList().erase(I);<br></pre>
878 <!--_______________________________________________________________________-->
879 <div class="doc_subsubsection">
880 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
884 <div class="doc_text">
886 <p><i>Replacing individual instructions</i></p>
888 <p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
889 permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
890 and <tt>ReplaceInstWithInst</tt>.</p>
892 <h4><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h4>
895 <li><tt>ReplaceInstWithValue</tt>
897 <p>This function replaces all uses (within a basic block) of a given
898 instruction with a value, and then removes the original instruction. The
899 following example illustrates the replacement of the result of a particular
900 <tt>AllocaInst</tt> that allocates memory for a single integer with an null
901 pointer to an integer.</p>
903 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithValue(instToReplace->getParent()->getInstList(), ii,<br> Constant::getNullValue(PointerType::get(Type::IntTy)));<br></pre></li>
905 <li><tt>ReplaceInstWithInst</tt>
907 <p>This function replaces a particular instruction with another
908 instruction. The following example illustrates the replacement of one
909 <tt>AllocaInst</tt> with another.</p>
911 <pre>AllocaInst* instToReplace = ...;<br>BasicBlock::iterator ii(instToReplace);<br>ReplaceInstWithInst(instToReplace->getParent()->getInstList(), ii,<br> new AllocaInst(Type::IntTy, 0, "ptrToReplacedInt"));<br></pre></li>
914 <p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
916 <p>You can use <tt>Value::replaceAllUsesWith</tt> and
917 <tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
918 doxygen documentation for the <a href="/doxygen/structllvm_1_1Value.html">Value Class</a>
919 and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
922 <!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
923 include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
924 ReplaceInstWithValue, ReplaceInstWithInst -->
928 <!-- *********************************************************************** -->
929 <div class="doc_section">
930 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
932 <!-- *********************************************************************** -->
934 <div class="doc_text">
936 <p>The Core LLVM classes are the primary means of representing the program
937 being inspected or transformed. The core LLVM classes are defined in
938 header files in the <tt>include/llvm/</tt> directory, and implemented in
939 the <tt>lib/VMCore</tt> directory.</p>
943 <!-- ======================================================================= -->
944 <div class="doc_subsection">
945 <a name="Value">The <tt>Value</tt> class</a>
950 <p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
952 doxygen info: <a href="/doxygen/structllvm_1_1Value.html">Value Class</a></p>
954 <p>The <tt>Value</tt> class is the most important class in the LLVM Source
955 base. It represents a typed value that may be used (among other things) as an
956 operand to an instruction. There are many different types of <tt>Value</tt>s,
957 such as <a href="#Constant"><tt>Constant</tt></a>s,<a
958 href="#Argument"><tt>Argument</tt></a>s. Even <a
959 href="#Instruction"><tt>Instruction</tt></a>s and <a
960 href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
962 <p>A particular <tt>Value</tt> may be used many times in the LLVM representation
963 for a program. For example, an incoming argument to a function (represented
964 with an instance of the <a href="#Argument">Argument</a> class) is "used" by
965 every instruction in the function that references the argument. To keep track
966 of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
967 href="#User"><tt>User</tt></a>s that is using it (the <a
968 href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
969 graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
970 def-use information in the program, and is accessible through the <tt>use_</tt>*
971 methods, shown below.</p>
973 <p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
974 and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
975 method. In addition, all LLVM values can be named. The "name" of the
976 <tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
978 <pre> %<b>foo</b> = add int 1, 2<br></pre>
980 <p><a name="#nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
981 that the name of any value may be missing (an empty string), so names should
982 <b>ONLY</b> be used for debugging (making the source code easier to read,
983 debugging printouts), they should not be used to keep track of values or map
984 between them. For this purpose, use a <tt>std::map</tt> of pointers to the
985 <tt>Value</tt> itself instead.</p>
987 <p>One important aspect of LLVM is that there is no distinction between an SSA
988 variable and the operation that produces it. Because of this, any reference to
989 the value produced by an instruction (or the value available as an incoming
990 argument, for example) is represented as a direct pointer to the instance of
992 represents this value. Although this may take some getting used to, it
993 simplifies the representation and makes it easier to manipulate.</p>
997 <!-- _______________________________________________________________________ -->
998 <div class="doc_subsubsection">
999 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
1002 <div class="doc_text">
1005 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
1007 <tt>Value::use_const_iterator</tt> - Typedef for const_iterator over
1009 <tt>unsigned use_size()</tt> - Returns the number of users of the
1011 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
1012 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
1014 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
1016 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
1017 element in the list.
1018 <p> These methods are the interface to access the def-use
1019 information in LLVM. As with all other iterators in LLVM, the naming
1020 conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
1022 <li><tt><a href="#Type">Type</a> *getType() const</tt>
1023 <p>This method returns the Type of the Value.</p>
1025 <li><tt>bool hasName() const</tt><br>
1026 <tt>std::string getName() const</tt><br>
1027 <tt>void setName(const std::string &Name)</tt>
1028 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
1029 be aware of the <a href="#nameWarning">precaution above</a>.</p>
1031 <li><tt>void replaceAllUsesWith(Value *V)</tt>
1033 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
1034 href="#User"><tt>User</tt>s</a> of the current value to refer to
1035 "<tt>V</tt>" instead. For example, if you detect that an instruction always
1036 produces a constant value (for example through constant folding), you can
1037 replace all uses of the instruction with the constant like this:</p>
1039 <pre> Inst->replaceAllUsesWith(ConstVal);<br></pre>
1044 <!-- ======================================================================= -->
1045 <div class="doc_subsection">
1046 <a name="User">The <tt>User</tt> class</a>
1049 <div class="doc_text">
1052 <tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
1053 doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
1054 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1056 <p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
1057 refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
1058 that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
1059 referring to. The <tt>User</tt> class itself is a subclass of
1062 <p>The operands of a <tt>User</tt> point directly to the LLVM <a
1063 href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
1064 Single Assignment (SSA) form, there can only be one definition referred to,
1065 allowing this direct connection. This connection provides the use-def
1066 information in LLVM.</p>
1070 <!-- _______________________________________________________________________ -->
1071 <div class="doc_subsubsection">
1072 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
1075 <div class="doc_text">
1077 <p>The <tt>User</tt> class exposes the operand list in two ways: through
1078 an index access interface and through an iterator based interface.</p>
1081 <li><tt>Value *getOperand(unsigned i)</tt><br>
1082 <tt>unsigned getNumOperands()</tt>
1083 <p> These two methods expose the operands of the <tt>User</tt> in a
1084 convenient form for direct access.</p></li>
1086 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
1088 <tt>User::op_const_iterator</tt> <tt>use_iterator op_begin()</tt> -
1089 Get an iterator to the start of the operand list.<br>
1090 <tt>use_iterator op_end()</tt> - Get an iterator to the end of the
1092 <p> Together, these methods make up the iterator based interface to
1093 the operands of a <tt>User</tt>.</p></li>
1098 <!-- ======================================================================= -->
1099 <div class="doc_subsection">
1100 <a name="Instruction">The <tt>Instruction</tt> class</a>
1103 <div class="doc_text">
1105 <p><tt>#include "</tt><tt><a
1106 href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
1107 doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
1108 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1109 href="#Value"><tt>Value</tt></a></p>
1111 <p>The <tt>Instruction</tt> class is the common base class for all LLVM
1112 instructions. It provides only a few methods, but is a very commonly used
1113 class. The primary data tracked by the <tt>Instruction</tt> class itself is the
1114 opcode (instruction type) and the parent <a
1115 href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
1116 into. To represent a specific type of instruction, one of many subclasses of
1117 <tt>Instruction</tt> are used.</p>
1119 <p> Because the <tt>Instruction</tt> class subclasses the <a
1120 href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
1121 way as for other <a href="#User"><tt>User</tt></a>s (with the
1122 <tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
1123 <tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
1124 the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
1125 file contains some meta-data about the various different types of instructions
1126 in LLVM. It describes the enum values that are used as opcodes (for example
1127 <tt>Instruction::Add</tt> and <tt>Instruction::SetLE</tt>), as well as the
1128 concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
1129 example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
1130 href="#SetCondInst">SetCondInst</a></tt>). Unfortunately, the use of macros in
1131 this file confuses doxygen, so these enum values don't show up correctly in the
1132 <a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
1136 <!-- _______________________________________________________________________ -->
1137 <div class="doc_subsubsection">
1138 <a name="m_Instruction">Important Public Members of the <tt>Instruction</tt>
1142 <div class="doc_text">
1145 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
1146 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
1147 this <tt>Instruction</tt> is embedded into.</p></li>
1148 <li><tt>bool mayWriteToMemory()</tt>
1149 <p>Returns true if the instruction writes to memory, i.e. it is a
1150 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
1151 <li><tt>unsigned getOpcode()</tt>
1152 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
1153 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
1154 <p>Returns another instance of the specified instruction, identical
1155 in all ways to the original except that the instruction has no parent
1156 (ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
1157 and it has no name</p></li>
1162 <!-- ======================================================================= -->
1163 <div class="doc_subsection">
1164 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
1167 <div class="doc_text">
1170 href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
1171 doxygen info: <a href="/doxygen/structllvm_1_1BasicBlock.html">BasicBlock
1173 Superclass: <a href="#Value"><tt>Value</tt></a></p>
1175 <p>This class represents a single entry multiple exit section of the code,
1176 commonly known as a basic block by the compiler community. The
1177 <tt>BasicBlock</tt> class maintains a list of <a
1178 href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
1179 Matching the language definition, the last element of this list of instructions
1180 is always a terminator instruction (a subclass of the <a
1181 href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
1183 <p>In addition to tracking the list of instructions that make up the block, the
1184 <tt>BasicBlock</tt> class also keeps track of the <a
1185 href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
1187 <p>Note that <tt>BasicBlock</tt>s themselves are <a
1188 href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
1189 like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
1194 <!-- _______________________________________________________________________ -->
1195 <div class="doc_subsubsection">
1196 <a name="m_BasicBlock">Important Public Members of the <tt>BasicBlock</tt>
1200 <div class="doc_text">
1204 <li><tt>BasicBlock(const std::string &Name = "", </tt><tt><a
1205 href="#Function">Function</a> *Parent = 0)</tt>
1207 <p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
1208 insertion into a function. The constructor optionally takes a name for the new
1209 block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
1210 the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
1211 automatically inserted at the end of the specified <a
1212 href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
1213 manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
1215 <li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
1216 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
1217 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1218 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt> -
1219 STL-style functions for accessing the instruction list.
1221 <p>These methods and typedefs are forwarding functions that have the same
1222 semantics as the standard library methods of the same names. These methods
1223 expose the underlying instruction list of a basic block in a way that is easy to
1224 manipulate. To get the full complement of container operations (including
1225 operations to update the list), you must use the <tt>getInstList()</tt>
1228 <li><tt>BasicBlock::InstListType &getInstList()</tt>
1230 <p>This method is used to get access to the underlying container that actually
1231 holds the Instructions. This method must be used when there isn't a forwarding
1232 function in the <tt>BasicBlock</tt> class for the operation that you would like
1233 to perform. Because there are no forwarding functions for "updating"
1234 operations, you need to use this if you want to update the contents of a
1235 <tt>BasicBlock</tt>.</p></li>
1237 <li><tt><a href="#Function">Function</a> *getParent()</tt>
1239 <p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
1240 embedded into, or a null pointer if it is homeless.</p></li>
1242 <li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
1244 <p> Returns a pointer to the terminator instruction that appears at the end of
1245 the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
1246 instruction in the block is not a terminator, then a null pointer is
1253 <!-- ======================================================================= -->
1254 <div class="doc_subsection">
1255 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
1258 <div class="doc_text">
1261 href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
1262 doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
1264 Superclasses: <a href="#User"><tt>User</tt></a>, <a
1265 href="#Value"><tt>Value</tt></a></p>
1267 <p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
1268 href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
1269 visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
1270 Because they are visible at global scope, they are also subject to linking with
1271 other globals defined in different translation units. To control the linking
1272 process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
1273 <tt>GlobalValue</tt>s know whether they have internal or external linkage, as
1274 defined by the <tt>LinkageTypes</tt> enumeration.</p>
1276 <p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
1277 <tt>static</tt> in C), it is not visible to code outside the current translation
1278 unit, and does not participate in linking. If it has external linkage, it is
1279 visible to external code, and does participate in linking. In addition to
1280 linkage information, <tt>GlobalValue</tt>s keep track of which <a
1281 href="#Module"><tt>Module</tt></a> they are currently part of.</p>
1283 <p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
1284 by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
1285 global is always a pointer to its contents. It is important to remember this
1286 when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
1287 be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
1288 subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
1289 int]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
1290 the address of the first element of this array and the value of the
1291 <tt>GlobalVariable</tt> are the same, they have different types. The
1292 <tt>GlobalVariable</tt>'s type is <tt>[24 x int]</tt>. The first element's type
1293 is <tt>int.</tt> Because of this, accessing a global value requires you to
1294 dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
1295 can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
1296 Language Reference Manual</a>.</p>
1300 <!-- _______________________________________________________________________ -->
1301 <div class="doc_subsubsection">
1302 <a name="m_GlobalValue">Important Public Members of the <tt>GlobalValue</tt>
1306 <div class="doc_text">
1309 <li><tt>bool hasInternalLinkage() const</tt><br>
1310 <tt>bool hasExternalLinkage() const</tt><br>
1311 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
1312 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
1315 <li><tt><a href="#Module">Module</a> *getParent()</tt>
1316 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
1317 GlobalValue is currently embedded into.</p></li>
1322 <!-- ======================================================================= -->
1323 <div class="doc_subsection">
1324 <a name="Function">The <tt>Function</tt> class</a>
1327 <div class="doc_text">
1330 href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
1331 info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
1332 Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1333 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1335 <p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
1336 actually one of the more complex classes in the LLVM heirarchy because it must
1337 keep track of a large amount of data. The <tt>Function</tt> class keeps track
1338 of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal <a
1339 href="#Argument"><tt>Argument</tt></a>s, and a <a
1340 href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
1342 <p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
1343 commonly used part of <tt>Function</tt> objects. The list imposes an implicit
1344 ordering of the blocks in the function, which indicate how the code will be
1345 layed out by the backend. Additionally, the first <a
1346 href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
1347 <tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
1348 block. There are no implicit exit nodes, and in fact there may be multiple exit
1349 nodes from a single <tt>Function</tt>. If the <a
1350 href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
1351 the <tt>Function</tt> is actually a function declaration: the actual body of the
1352 function hasn't been linked in yet.</p>
1354 <p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
1355 <tt>Function</tt> class also keeps track of the list of formal <a
1356 href="#Argument"><tt>Argument</tt></a>s that the function receives. This
1357 container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
1358 nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
1359 the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
1361 <p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
1362 LLVM feature that is only used when you have to look up a value by name. Aside
1363 from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
1364 internally to make sure that there are not conflicts between the names of <a
1365 href="#Instruction"><tt>Instruction</tt></a>s, <a
1366 href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
1367 href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
1369 <p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
1370 and therefore also a <a href="#Constant">Constant</a>. The value of the function
1371 is its address (after linking) which is guaranteed to be constant.</p>
1374 <!-- _______________________________________________________________________ -->
1375 <div class="doc_subsubsection">
1376 <a name="m_Function">Important Public Members of the <tt>Function</tt>
1380 <div class="doc_text">
1383 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
1384 *Ty, LinkageTypes Linkage, const std::string &N = "", Module* Parent = 0)</tt>
1386 <p>Constructor used when you need to create new <tt>Function</tt>s to add
1387 the the program. The constructor must specify the type of the function to
1388 create and what type of linkage the function should have. The <a
1389 href="#FunctionType"><tt>FunctionType</tt></a> argument
1390 specifies the formal arguments and return value for the function. The same
1391 <a href="#FunctionTypel"><tt>FunctionType</tt></a> value can be used to
1392 create multiple functions. The <tt>Parent</tt> argument specifies the Module
1393 in which the function is defined. If this argument is provided, the function
1394 will automatically be inserted into that module's list of
1397 <li><tt>bool isExternal()</tt>
1399 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
1400 function is "external", it does not have a body, and thus must be resolved
1401 by linking with a function defined in a different translation unit.</p></li>
1403 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
1404 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
1406 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1407 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1409 <p>These are forwarding methods that make it easy to access the contents of
1410 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
1413 <li><tt>Function::BasicBlockListType &getBasicBlockList()</tt>
1415 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
1416 is necessary to use when you need to update the list or perform a complex
1417 action that doesn't have a forwarding method.</p></li>
1419 <li><tt>Function::aiterator</tt> - Typedef for the argument list
1421 <tt>Function::const_aiterator</tt> - Typedef for const_iterator.<br>
1423 <tt>abegin()</tt>, <tt>aend()</tt>, <tt>afront()</tt>, <tt>aback()</tt>,
1424 <tt>asize()</tt>, <tt>aempty()</tt>, <tt>arbegin()</tt>, <tt>arend()</tt>
1426 <p>These are forwarding methods that make it easy to access the contents of
1427 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
1430 <li><tt>Function::ArgumentListType &getArgumentList()</tt>
1432 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
1433 necessary to use when you need to update the list or perform a complex
1434 action that doesn't have a forwarding method.</p></li>
1436 <li><tt><a href="#BasicBlock">BasicBlock</a> &getEntryBlock()</tt>
1438 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
1439 function. Because the entry block for the function is always the first
1440 block, this returns the first block of the <tt>Function</tt>.</p></li>
1442 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
1443 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
1445 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
1446 <tt>Function</tt> and returns the return type of the function, or the <a
1447 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
1450 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1452 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1453 for this <tt>Function</tt>.</p></li>
1458 <!-- ======================================================================= -->
1459 <div class="doc_subsection">
1460 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
1463 <div class="doc_text">
1466 href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
1468 doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
1469 Class</a><br> Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>, <a
1470 href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
1472 <p>Global variables are represented with the (suprise suprise)
1473 <tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
1474 subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
1475 always referenced by their address (global values must live in memory, so their
1476 "name" refers to their address). See <a
1477 href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global variables
1478 may have an initial value (which must be a <a
1479 href="#Constant"><tt>Constant</tt></a>), and if they have an initializer, they
1480 may be marked as "constant" themselves (indicating that their contents never
1481 change at runtime).</p>
1485 <!-- _______________________________________________________________________ -->
1486 <div class="doc_subsubsection">
1487 <a name="m_GlobalVariable">Important Public Members of the
1488 <tt>GlobalVariable</tt> class</a>
1491 <div class="doc_text">
1494 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
1495 isConstant, LinkageTypes& Linkage, <a href="#Constant">Constant</a>
1496 *Initializer = 0, const std::string &Name = "", Module* Parent = 0)</tt>
1498 <p>Create a new global variable of the specified type. If
1499 <tt>isConstant</tt> is true then the global variable will be marked as
1500 unchanging for the program. The Linkage parameter specifies the type of
1501 linkage (internal, external, weak, linkonce, appending) for the variable. If
1502 the linkage is InternalLinkage, WeakLinkage, or LinkOnceLinkage, then
1503 the resultant global variable will have internal linkage. AppendingLinkage
1504 concatenates together all instances (in different translation units) of the
1505 variable into a single variable but is only applicable to arrays. See
1506 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
1507 further details on linkage types. Optionally an initializer, a name, and the
1508 module to put the variable into may be specified for the global variable as
1511 <li><tt>bool isConstant() const</tt>
1513 <p>Returns true if this is a global variable that is known not to
1514 be modified at runtime.</p></li>
1516 <li><tt>bool hasInitializer()</tt>
1518 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
1520 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
1522 <p>Returns the intial value for a <tt>GlobalVariable</tt>. It is not legal
1523 to call this method if there is no initializer.</p></li>
1528 <!-- ======================================================================= -->
1529 <div class="doc_subsection">
1530 <a name="Module">The <tt>Module</tt> class</a>
1533 <div class="doc_text">
1536 href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
1537 <a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
1539 <p>The <tt>Module</tt> class represents the top level structure present in LLVM
1540 programs. An LLVM module is effectively either a translation unit of the
1541 original program or a combination of several translation units merged by the
1542 linker. The <tt>Module</tt> class keeps track of a list of <a
1543 href="#Function"><tt>Function</tt></a>s, a list of <a
1544 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
1545 href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
1546 helpful member functions that try to make common operations easy.</p>
1550 <!-- _______________________________________________________________________ -->
1551 <div class="doc_subsubsection">
1552 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
1555 <div class="doc_text">
1558 <li><tt>Module::Module(std::string name = "")</tt></li>
1561 <p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
1562 provide a name for it (probably based on the name of the translation unit).</p>
1565 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
1566 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
1568 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
1569 <tt>size()</tt>, <tt>empty()</tt>, <tt>rbegin()</tt>, <tt>rend()</tt>
1571 <p>These are forwarding methods that make it easy to access the contents of
1572 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
1575 <li><tt>Module::FunctionListType &getFunctionList()</tt>
1577 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
1578 necessary to use when you need to update the list or perform a complex
1579 action that doesn't have a forwarding method.</p>
1581 <p><!-- Global Variable --></p></li>
1587 <li><tt>Module::giterator</tt> - Typedef for global variable list iterator<br>
1589 <tt>Module::const_giterator</tt> - Typedef for const_iterator.<br>
1591 <tt>gbegin()</tt>, <tt>gend()</tt>, <tt>gfront()</tt>, <tt>gback()</tt>,
1592 <tt>gsize()</tt>, <tt>gempty()</tt>, <tt>grbegin()</tt>, <tt>grend()</tt>
1594 <p> These are forwarding methods that make it easy to access the contents of
1595 a <tt>Module</tt> object's <a
1596 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
1598 <li><tt>Module::GlobalListType &getGlobalList()</tt>
1600 <p>Returns the list of <a
1601 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
1602 use when you need to update the list or perform a complex action that
1603 doesn't have a forwarding method.</p>
1605 <p><!-- Symbol table stuff --> </p></li>
1611 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
1613 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1614 for this <tt>Module</tt>.</p>
1616 <p><!-- Convenience methods --></p></li>
1622 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
1623 &Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
1625 <p>Look up the specified function in the <tt>Module</tt> <a
1626 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
1627 <tt>null</tt>.</p></li>
1629 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
1630 std::string &Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
1632 <p>Look up the specified function in the <tt>Module</tt> <a
1633 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
1634 external declaration for the function and return it.</p></li>
1636 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
1638 <p>If there is at least one entry in the <a
1639 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
1640 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
1643 <li><tt>bool addTypeName(const std::string &Name, const <a
1644 href="#Type">Type</a> *Ty)</tt>
1646 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
1647 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
1648 name, true is returned and the <a
1649 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
1654 <!-- ======================================================================= -->
1655 <div class="doc_subsection">
1656 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
1659 <div class="doc_text">
1661 <p>Constant represents a base class for different types of constants. It
1662 is subclassed by ConstantBool, ConstantInt, ConstantSInt, ConstantUInt,
1663 ConstantArray etc for representing the various types of Constants.</p>
1667 <!-- _______________________________________________________________________ -->
1668 <div class="doc_subsubsection">
1669 <a name="m_Value">Important Public Methods</a>
1672 <div class="doc_text">
1675 <hr> Important Subclasses of Constant
1678 <li>ConstantSInt : This subclass of Constant represents a signed
1681 <li><tt>int64_t getValue() const</tt>: Returns the underlying value of
1682 this constant. </li>
1685 <li>ConstantUInt : This class represents an unsigned integer.
1687 <li><tt>uint64_t getValue() const</tt>: Returns the underlying value
1688 of this constant. </li>
1691 <li>ConstantFP : This class represents a floating point constant.
1693 <li><tt>double getValue() const</tt>: Returns the underlying value of
1694 this constant. </li>
1697 <li>ConstantBool : This represents a boolean constant.
1699 <li><tt>bool getValue() const</tt>: Returns the underlying value of
1700 this constant. </li>
1703 <li>ConstantArray : This represents a constant array.
1705 <li><tt>const std::vector<Use> &getValues() const</tt>:
1706 Returns a Vecotr of component constants that makeup this array. </li>
1709 <li>ConstantStruct : This represents a constant struct.
1711 <li><tt>const std::vector<Use> &getValues() const</tt>:
1712 Returns a Vecotr of component constants that makeup this array. </li>
1715 <li>GlobalValue : This represents either a global variable or a
1716 function. In either case, the value is a constant fixed address
1724 <!-- ======================================================================= -->
1725 <div class="doc_subsection">
1726 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
1729 <div class="doc_text">
1731 <p>Type as noted earlier is also a subclass of a Value class. Any primitive
1732 type (like int, short etc) in LLVM is an instance of Type Class. All other
1733 types are instances of subclasses of type like FunctionType, ArrayType
1734 etc. DerivedType is the interface for all such dervied types including
1735 FunctionType, ArrayType, PointerType, StructType. Types can have names. They can
1736 be recursive (StructType). There exists exactly one instance of any type
1737 structure at a time. This allows using pointer equality of Type *s for comparing
1742 <!-- _______________________________________________________________________ -->
1743 <div class="doc_subsubsection">
1744 <a name="m_Value">Important Public Methods</a>
1747 <div class="doc_text">
1751 <li><tt>bool isSigned() const</tt>: Returns whether an integral numeric type
1752 is signed. This is true for SByteTy, ShortTy, IntTy, LongTy. Note that this is
1753 not true for Float and Double. </li>
1755 <li><tt>bool isUnsigned() const</tt>: Returns whether a numeric type is
1756 unsigned. This is not quite the complement of isSigned... nonnumeric types
1757 return false as they do with isSigned. This returns true for UByteTy,
1758 UShortTy, UIntTy, and ULongTy. </li>
1760 <li><tt>bool isInteger() const</tt>: Equivalent to isSigned() || isUnsigned().</li>
1762 <li><tt>bool isIntegral() const</tt>: Returns true if this is an integral
1763 type, which is either Bool type or one of the Integer types.</li>
1765 <li><tt>bool isFloatingPoint()</tt>: Return true if this is one of the two
1766 floating point types.</li>
1768 <li><tt>isLosslesslyConvertableTo (const Type *Ty) const</tt>: Return true if
1769 this type can be converted to 'Ty' without any reinterpretation of bits. For
1770 example, uint to int or one pointer type to another.</li>
1773 <p>Derived Types</p>
1776 <li>SequentialType : This is subclassed by ArrayType and PointerType
1778 <li><tt>const Type * getElementType() const</tt>: Returns the type of
1779 each of the elements in the sequential type. </li>
1782 <li>ArrayType : This is a subclass of SequentialType and defines
1783 interface for array types.
1785 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
1786 elements in the array. </li>
1789 <li>PointerType : Subclass of SequentialType for pointer types. </li>
1790 <li>StructType : subclass of DerivedTypes for struct types </li>
1791 <li>FunctionType : subclass of DerivedTypes for function types.
1793 <li><tt>bool isVarArg() const</tt>: Returns true if its a vararg
1795 <li><tt> const Type * getReturnType() const</tt>: Returns the
1796 return type of the function.</li>
1797 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
1798 the type of the ith parameter.</li>
1799 <li><tt> const unsigned getNumParams() const</tt>: Returns the
1800 number of formal parameters.</li>
1809 <!-- ======================================================================= -->
1810 <div class="doc_subsection">
1811 <a name="Argument">The <tt>Argument</tt> class</a>
1814 <div class="doc_text">
1816 <p>This subclass of Value defines the interface for incoming formal
1817 arguments to a function. A Function maitanis a list of its formal
1818 arguments. An argument has a pointer to the parent Function.</p>
1822 <!-- ======================================================================= -->
1823 <div class="doc_subsection">
1824 <a name="SymbolTable">The <tt>SymbolTable</tt> class</a>
1826 <div class="doc_text">
1827 <p>This class provides a symbol table that the
1828 <a href="#Function"><tt>Function</tt></a> and <a href="#Module">
1829 <tt>Module</tt></a> classes use for naming definitions. The symbol table can
1830 provide a name for any <a href="#Value"><tt>Value</tt></a> or
1831 <a href="#Type"><tt>Type</tt></a>. <tt>SymbolTable</tt> is an abstract data
1832 type. It hides the data it contains and provides access to it through a
1833 controlled interface.</p>
1835 <p>To use the <tt>SymbolTable</tt> well, you need to understand the
1836 structure of the information it holds. The class contains two
1837 <tt>std::map</tt> objects. The first, <tt>pmap</tt>, is a map of
1838 <tt>Type*</tt> to maps of name (<tt>std::string</tt>) to <tt>Value*</tt>.
1839 The second, <tt>tmap</tt>, is a map of names to <tt>Type*</tt>. Thus, Values
1840 are stored in two-dimensions and accessed by <tt>Type</tt> and name. Types,
1841 however, are stored in a single dimension and accessed only by name.</p>
1843 <p>The interface of this class provides three basic types of operations:
1845 <li><em>Accessors</em>. Accessors provide read-only access to information
1846 such as finding a value for a name with the
1847 <a href="#SymbolTable_lookup">lookup</a> method.</li>
1848 <li><em>Mutators</em>. Mutators allow the user to add information to the
1849 <tt>SymbolTable</tt> with methods like
1850 <a href="#SymbolTable_insert"><tt>insert</tt></a>.</li>
1851 <li><em>Iterators</em>. Iterators allow the user to traverse the content
1852 of the symbol table in well defined ways, such as the method
1853 <a href="#SymbolTable_type_begin"><tt>type_begin</tt></a>.</li>
1858 <dt><tt>Value* lookup(const Type* Ty, const std::string& name) const</tt>:
1860 <dd>The <tt>lookup</tt> method searches the type plane given by the
1861 <tt>Ty</tt> parameter for a <tt>Value</tt> with the provided <tt>name</tt>.
1862 If a suitable <tt>Value</tt> is not found, null is returned.</dd>
1864 <dt><tt>Type* lookupType( const std::string& name) const</tt>:</dt>
1865 <dd>The <tt>lookupType</tt> method searches through the types for a
1866 <tt>Type</tt> with the provided <tt>name</tt>. If a suitable <tt>Type</tt>
1867 is not found, null is returned.</dd>
1869 <dt><tt>bool hasTypes() const</tt>:</dt>
1870 <dd>This function returns true if an entry has been made into the type
1873 <dt><tt>bool isEmpty() const</tt>:</dt>
1874 <dd>This function returns true if both the value and types maps are
1877 <dt><tt>std::string get_name(const Value*) const</tt>:</dt>
1878 <dd>This function returns the name of the Value provided or the empty
1879 string if the Value is not in the symbol table.</dd>
1881 <dt><tt>std::string get_name(const Type*) const</tt>:</dt>
1882 <dd>This function returns the name of the Type provided or the empty
1883 string if the Type is not in the symbol table.</dd>
1888 <dt><tt>void insert(Value *Val)</tt>:</dt>
1889 <dd>This method adds the provided value to the symbol table. The Value must
1890 have both a name and a type which are extracted and used to place the value
1891 in the correct type plane under the value's name.</dd>
1893 <dt><tt>void insert(const std::string& Name, Value *Val)</tt>:</dt>
1894 <dd> Inserts a constant or type into the symbol table with the specified
1895 name. There can be a many to one mapping between names and constants
1898 <dt><tt>void insert(const std::string& Name, Type *Typ)</tt>:</dt>
1899 <dd> Inserts a type into the symbol table with the specified name. There
1900 can be a many-to-one mapping between names and types. This method
1901 allows a type with an existing entry in the symbol table to get
1904 <dt><tt>void remove(Value* Val)</tt>:</dt>
1905 <dd> This method removes a named value from the symbol table. The
1906 type and name of the Value are extracted from \p N and used to
1907 lookup the Value in the correct type plane. If the Value is
1908 not in the symbol table, this method silently ignores the
1911 <dt><tt>void remove(Type* Typ)</tt>:</dt>
1912 <dd> This method removes a named type from the symbol table. The
1913 name of the type is extracted from \P T and used to look up
1914 the Type in the type map. If the Type is not in the symbol
1915 table, this method silently ignores the request.</dd>
1917 <dt><tt>Value* remove(const std::string& Name, Value *Val)</tt>:</dt>
1918 <dd> Remove a constant or type with the specified name from the
1921 <dt><tt>Type* remove(const std::string& Name, Type* T)</tt>:</dt>
1922 <dd> Remove a type with the specified name from the symbol table.
1923 Returns the removed Type.</dd>
1925 <dt><tt>Value *value_remove(const value_iterator& It)</tt>:</dt>
1926 <dd> Removes a specific value from the symbol table.
1927 Returns the removed value.</dd>
1929 <dt><tt>bool strip()</tt>:</dt>
1930 <dd> This method will strip the symbol table of its names leaving
1931 the type and values. </dd>
1933 <dt><tt>void clear()</tt>:</dt>
1934 <dd>Empty the symbol table completely.</dd>
1938 <p>The following functions describe three types of iterators you can obtain
1939 the beginning or end of the sequence for both const and non-const. It is
1940 important to keep track of the different kinds of iterators. There are
1941 three idioms worth pointing out:</p>
1943 <tr><th>Units</th><th>Iterator</th><th>Idiom</th></tr>
1945 <td align="left">Planes Of name/Value maps</td><td>PI</td>
1946 <td align="left"><tt><pre>
1947 for (SymbolTable::plane_const_iterator PI = ST.plane_begin(),
1948 PE = ST.plane_end(); PI != PE; ++PI ) {
1949 PI->first // This is the Type* of the plane
1950 PI->second // This is the SymbolTable::ValueMap of name/Value pairs
1954 <td align="left">All name/Type Pairs</td><td>TI</td>
1955 <td align="left"><tt><pre>
1956 for (SymbolTable::type_const_iterator TI = ST.type_begin(),
1957 TE = ST.type_end(); TI != TE; ++TI )
1958 TI->first // This is the name of the type
1959 TI->second // This is the Type* value associated with the name
1963 <td align="left">name/Value pairs in a plane</td><td>VI</td>
1964 <td align="left"><tt><pre>
1965 for (SymbolTable::value_const_iterator VI = ST.value_begin(SomeType),
1966 VE = ST.value_end(SomeType); VI != VE; ++VI )
1967 VI->first // This is the name of the Value
1968 VI->second // This is the Value* value associated with the name
1972 <p>Using the recommended iterator names and idioms will help you avoid
1973 making mistakes. Of particular note, make sure that whenever you use
1974 value_begin(SomeType) that you always compare the resulting iterator
1975 with value_end(SomeType) not value_end(SomeOtherType) or else you
1976 will loop infinitely.</p>
1980 <dt><tt>plane_iterator plane_begin()</tt>:</dt>
1981 <dd>Get an iterator that starts at the beginning of the type planes.
1982 The iterator will iterate over the Type/ValueMap pairs in the
1985 <dt><tt>plane_const_iterator plane_begin() const</tt>:</dt>
1986 <dd>Get a const_iterator that starts at the beginning of the type
1987 planes. The iterator will iterate over the Type/ValueMap pairs
1988 in the type planes. </dd>
1990 <dt><tt>plane_iterator plane_end()</tt>:</dt>
1991 <dd>Get an iterator at the end of the type planes. This serves as
1992 the marker for end of iteration over the type planes.</dd>
1994 <dt><tt>plane_const_iterator plane_end() const</tt>:</dt>
1995 <dd>Get a const_iterator at the end of the type planes. This serves as
1996 the marker for end of iteration over the type planes.</dd>
1998 <dt><tt>value_iterator value_begin(const Type *Typ)</tt>:</dt>
1999 <dd>Get an iterator that starts at the beginning of a type plane.
2000 The iterator will iterate over the name/value pairs in the type plane.
2001 Note: The type plane must already exist before using this.</dd>
2003 <dt><tt>value_const_iterator value_begin(const Type *Typ) const</tt>:</dt>
2004 <dd>Get a const_iterator that starts at the beginning of a type plane.
2005 The iterator will iterate over the name/value pairs in the type plane.
2006 Note: The type plane must already exist before using this.</dd>
2008 <dt><tt>value_iterator value_end(const Type *Typ)</tt>:</dt>
2009 <dd>Get an iterator to the end of a type plane. This serves as the marker
2010 for end of iteration of the type plane.
2011 Note: The type plane must already exist before using this.</dd>
2013 <dt><tt>value_const_iterator value_end(const Type *Typ) const</tt>:</dt>
2014 <dd>Get a const_iterator to the end of a type plane. This serves as the
2015 marker for end of iteration of the type plane.
2016 Note: the type plane must already exist before using this.</dd>
2018 <dt><tt>type_iterator type_begin()</tt>:</dt>
2019 <dd>Get an iterator to the start of the name/Type map.</dd>
2021 <dt><tt>type_const_iterator type_begin() cons</tt>:</dt>
2022 <dd> Get a const_iterator to the start of the name/Type map.</dd>
2024 <dt><tt>type_iterator type_end()</tt>:</dt>
2025 <dd>Get an iterator to the end of the name/Type map. This serves as the
2026 marker for end of iteration of the types.</dd>
2028 <dt><tt>type_const_iterator type_end() const</tt>:</dt>
2029 <dd>Get a const-iterator to the end of the name/Type map. This serves
2030 as the marker for end of iteration of the types.</dd>
2032 <dt><tt>plane_const_iterator find(const Type* Typ ) const</tt>:</dt>
2033 <dd>This method returns a plane_const_iterator for iteration over
2034 the type planes starting at a specific plane, given by \p Ty.</dd>
2036 <dt><tt>plane_iterator find( const Type* Typ </tt>:</dt>
2037 <dd>This method returns a plane_iterator for iteration over the
2038 type planes starting at a specific plane, given by \p Ty.</dd>
2040 <dt><tt>const ValueMap* findPlane( const Type* Typ ) cons</tt>:</dt>
2041 <dd>This method returns a ValueMap* for a specific type plane. This
2042 interface is deprecated and may go away in the future.</dd>
2046 <!-- *********************************************************************** -->
2049 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
2050 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
2051 <a href="http://validator.w3.org/check/referer"><img
2052 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
2054 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
2055 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
2056 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
2057 Last modified: $Date$