1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <meta http-equiv="Content-type" content="text/html;charset=UTF-8">
6 <title>LLVM Programmer's Manual</title>
7 <link rel="stylesheet" href="llvm.css" type="text/css">
12 LLVM Programmer's Manual
16 <li><a href="#introduction">Introduction</a></li>
17 <li><a href="#general">General Information</a>
19 <li><a href="#stl">The C++ Standard Template Library</a></li>
21 <li>The <tt>-time-passes</tt> option</li>
22 <li>How to use the LLVM Makefile system</li>
23 <li>How to write a regression test</li>
28 <li><a href="#apis">Important and useful LLVM APIs</a>
30 <li><a href="#isa">The <tt>isa<></tt>, <tt>cast<></tt>
31 and <tt>dyn_cast<></tt> templates</a> </li>
32 <li><a href="#string_apis">Passing strings (the <tt>StringRef</tt>
33 and <tt>Twine</tt> classes)</a>
35 <li><a href="#StringRef">The <tt>StringRef</tt> class</a> </li>
36 <li><a href="#Twine">The <tt>Twine</tt> class</a> </li>
39 <li><a href="#DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt>
42 <li><a href="#DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt>
43 and the <tt>-debug-only</tt> option</a> </li>
46 <li><a href="#Statistic">The <tt>Statistic</tt> class & <tt>-stats</tt>
49 <li>The <tt>InstVisitor</tt> template
50 <li>The general graph API
52 <li><a href="#ViewGraph">Viewing graphs while debugging code</a></li>
55 <li><a href="#datastructure">Picking the Right Data Structure for a Task</a>
57 <li><a href="#ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
59 <li><a href="#dss_arrayref">llvm/ADT/ArrayRef.h</a></li>
60 <li><a href="#dss_fixedarrays">Fixed Size Arrays</a></li>
61 <li><a href="#dss_heaparrays">Heap Allocated Arrays</a></li>
62 <li><a href="#dss_tinyptrvector">"llvm/ADT/TinyPtrVector.h"</a></li>
63 <li><a href="#dss_smallvector">"llvm/ADT/SmallVector.h"</a></li>
64 <li><a href="#dss_vector"><vector></a></li>
65 <li><a href="#dss_deque"><deque></a></li>
66 <li><a href="#dss_list"><list></a></li>
67 <li><a href="#dss_ilist">llvm/ADT/ilist.h</a></li>
68 <li><a href="#dss_packedvector">llvm/ADT/PackedVector.h</a></li>
69 <li><a href="#dss_other">Other Sequential Container Options</a></li>
71 <li><a href="#ds_string">String-like containers</a>
73 <li><a href="#dss_stringref">llvm/ADT/StringRef.h</a></li>
74 <li><a href="#dss_twine">llvm/ADT/Twine.h</a></li>
75 <li><a href="#dss_smallstring">llvm/ADT/SmallString.h</a></li>
76 <li><a href="#dss_stdstring">std::string</a></li>
78 <li><a href="#ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
80 <li><a href="#dss_sortedvectorset">A sorted 'vector'</a></li>
81 <li><a href="#dss_smallset">"llvm/ADT/SmallSet.h"</a></li>
82 <li><a href="#dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a></li>
83 <li><a href="#dss_denseset">"llvm/ADT/DenseSet.h"</a></li>
84 <li><a href="#dss_sparseset">"llvm/ADT/SparseSet.h"</a></li>
85 <li><a href="#dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a></li>
86 <li><a href="#dss_set"><set></a></li>
87 <li><a href="#dss_setvector">"llvm/ADT/SetVector.h"</a></li>
88 <li><a href="#dss_uniquevector">"llvm/ADT/UniqueVector.h"</a></li>
89 <li><a href="#dss_immutableset">"llvm/ADT/ImmutableSet.h"</a></li>
90 <li><a href="#dss_otherset">Other Set-Like Container Options</a></li>
92 <li><a href="#ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
94 <li><a href="#dss_sortedvectormap">A sorted 'vector'</a></li>
95 <li><a href="#dss_stringmap">"llvm/ADT/StringMap.h"</a></li>
96 <li><a href="#dss_indexedmap">"llvm/ADT/IndexedMap.h"</a></li>
97 <li><a href="#dss_densemap">"llvm/ADT/DenseMap.h"</a></li>
98 <li><a href="#dss_valuemap">"llvm/ADT/ValueMap.h"</a></li>
99 <li><a href="#dss_intervalmap">"llvm/ADT/IntervalMap.h"</a></li>
100 <li><a href="#dss_map"><map></a></li>
101 <li><a href="#dss_inteqclasses">"llvm/ADT/IntEqClasses.h"</a></li>
102 <li><a href="#dss_immutablemap">"llvm/ADT/ImmutableMap.h"</a></li>
103 <li><a href="#dss_othermap">Other Map-Like Container Options</a></li>
105 <li><a href="#ds_bit">BitVector-like containers</a>
107 <li><a href="#dss_bitvector">A dense bitvector</a></li>
108 <li><a href="#dss_smallbitvector">A "small" dense bitvector</a></li>
109 <li><a href="#dss_sparsebitvector">A sparse bitvector</a></li>
113 <li><a href="#common">Helpful Hints for Common Operations</a>
115 <li><a href="#inspection">Basic Inspection and Traversal Routines</a>
117 <li><a href="#iterate_function">Iterating over the <tt>BasicBlock</tt>s
118 in a <tt>Function</tt></a> </li>
119 <li><a href="#iterate_basicblock">Iterating over the <tt>Instruction</tt>s
120 in a <tt>BasicBlock</tt></a> </li>
121 <li><a href="#iterate_institer">Iterating over the <tt>Instruction</tt>s
122 in a <tt>Function</tt></a> </li>
123 <li><a href="#iterate_convert">Turning an iterator into a
124 class pointer</a> </li>
125 <li><a href="#iterate_complex">Finding call sites: a more
126 complex example</a> </li>
127 <li><a href="#calls_and_invokes">Treating calls and invokes
128 the same way</a> </li>
129 <li><a href="#iterate_chains">Iterating over def-use &
130 use-def chains</a> </li>
131 <li><a href="#iterate_preds">Iterating over predecessors &
132 successors of blocks</a></li>
135 <li><a href="#simplechanges">Making simple changes</a>
137 <li><a href="#schanges_creating">Creating and inserting new
138 <tt>Instruction</tt>s</a> </li>
139 <li><a href="#schanges_deleting">Deleting <tt>Instruction</tt>s</a> </li>
140 <li><a href="#schanges_replacing">Replacing an <tt>Instruction</tt>
141 with another <tt>Value</tt></a> </li>
142 <li><a href="#schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a> </li>
145 <li><a href="#create_types">How to Create Types</a></li>
147 <li>Working with the Control Flow Graph
149 <li>Accessing predecessors and successors of a <tt>BasicBlock</tt>
157 <li><a href="#threading">Threads and LLVM</a>
159 <li><a href="#startmultithreaded">Entering and Exiting Multithreaded Mode
161 <li><a href="#shutdown">Ending execution with <tt>llvm_shutdown()</tt></a></li>
162 <li><a href="#managedstatic">Lazy initialization with <tt>ManagedStatic</tt></a></li>
163 <li><a href="#llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a></li>
164 <li><a href="#jitthreading">Threads and the JIT</a></li>
168 <li><a href="#advanced">Advanced Topics</a>
171 <li><a href="#SymbolTable">The <tt>ValueSymbolTable</tt> class</a></li>
172 <li><a href="#UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a></li>
175 <li><a href="#coreclasses">The Core LLVM Class Hierarchy Reference</a>
177 <li><a href="#Type">The <tt>Type</tt> class</a> </li>
178 <li><a href="#Module">The <tt>Module</tt> class</a></li>
179 <li><a href="#Value">The <tt>Value</tt> class</a>
181 <li><a href="#User">The <tt>User</tt> class</a>
183 <li><a href="#Instruction">The <tt>Instruction</tt> class</a></li>
184 <li><a href="#Constant">The <tt>Constant</tt> class</a>
186 <li><a href="#GlobalValue">The <tt>GlobalValue</tt> class</a>
188 <li><a href="#Function">The <tt>Function</tt> class</a></li>
189 <li><a href="#GlobalVariable">The <tt>GlobalVariable</tt> class</a></li>
196 <li><a href="#BasicBlock">The <tt>BasicBlock</tt> class</a></li>
197 <li><a href="#Argument">The <tt>Argument</tt> class</a></li>
204 <div class="doc_author">
205 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>,
206 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a>,
207 <a href="mailto:ggreif@gmail.com">Gabor Greif</a>,
208 <a href="mailto:jstanley@cs.uiuc.edu">Joel Stanley</a>,
209 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> and
210 <a href="mailto:owen@apple.com">Owen Anderson</a></p>
213 <!-- *********************************************************************** -->
215 <a name="introduction">Introduction </a>
217 <!-- *********************************************************************** -->
221 <p>This document is meant to highlight some of the important classes and
222 interfaces available in the LLVM source-base. This manual is not
223 intended to explain what LLVM is, how it works, and what LLVM code looks
224 like. It assumes that you know the basics of LLVM and are interested
225 in writing transformations or otherwise analyzing or manipulating the
228 <p>This document should get you oriented so that you can find your
229 way in the continuously growing source code that makes up the LLVM
230 infrastructure. Note that this manual is not intended to serve as a
231 replacement for reading the source code, so if you think there should be
232 a method in one of these classes to do something, but it's not listed,
233 check the source. Links to the <a href="/doxygen/">doxygen</a> sources
234 are provided to make this as easy as possible.</p>
236 <p>The first section of this document describes general information that is
237 useful to know when working in the LLVM infrastructure, and the second describes
238 the Core LLVM classes. In the future this manual will be extended with
239 information describing how to use extension libraries, such as dominator
240 information, CFG traversal routines, and useful utilities like the <tt><a
241 href="/doxygen/InstVisitor_8h-source.html">InstVisitor</a></tt> template.</p>
245 <!-- *********************************************************************** -->
247 <a name="general">General Information</a>
249 <!-- *********************************************************************** -->
253 <p>This section contains general information that is useful if you are working
254 in the LLVM source-base, but that isn't specific to any particular API.</p>
256 <!-- ======================================================================= -->
258 <a name="stl">The C++ Standard Template Library</a>
263 <p>LLVM makes heavy use of the C++ Standard Template Library (STL),
264 perhaps much more than you are used to, or have seen before. Because of
265 this, you might want to do a little background reading in the
266 techniques used and capabilities of the library. There are many good
267 pages that discuss the STL, and several books on the subject that you
268 can get, so it will not be discussed in this document.</p>
270 <p>Here are some useful links:</p>
274 <li><a href="http://www.dinkumware.com/manuals/#Standard C++ Library">Dinkumware
275 C++ Library reference</a> - an excellent reference for the STL and other parts
276 of the standard C++ library.</li>
278 <li><a href="http://www.tempest-sw.com/cpp/">C++ In a Nutshell</a> - This is an
279 O'Reilly book in the making. It has a decent Standard Library
280 Reference that rivals Dinkumware's, and is unfortunately no longer free since the
281 book has been published.</li>
283 <li><a href="http://www.parashift.com/c++-faq-lite/">C++ Frequently Asked
286 <li><a href="http://www.sgi.com/tech/stl/">SGI's STL Programmer's Guide</a> -
288 href="http://www.sgi.com/tech/stl/stl_introduction.html">Introduction to the
291 <li><a href="http://www.research.att.com/%7Ebs/C++.html">Bjarne Stroustrup's C++
294 <li><a href="http://64.78.49.204/">
295 Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0 (even better, get
300 <p>You are also encouraged to take a look at the <a
301 href="CodingStandards.html">LLVM Coding Standards</a> guide which focuses on how
302 to write maintainable code more than where to put your curly braces.</p>
306 <!-- ======================================================================= -->
308 <a name="stl">Other useful references</a>
314 <li><a href="http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html">Using
315 static and shared libraries across platforms</a></li>
322 <!-- *********************************************************************** -->
324 <a name="apis">Important and useful LLVM APIs</a>
326 <!-- *********************************************************************** -->
330 <p>Here we highlight some LLVM APIs that are generally useful and good to
331 know about when writing transformations.</p>
333 <!-- ======================================================================= -->
335 <a name="isa">The <tt>isa<></tt>, <tt>cast<></tt> and
336 <tt>dyn_cast<></tt> templates</a>
341 <p>The LLVM source-base makes extensive use of a custom form of RTTI.
342 These templates have many similarities to the C++ <tt>dynamic_cast<></tt>
343 operator, but they don't have some drawbacks (primarily stemming from
344 the fact that <tt>dynamic_cast<></tt> only works on classes that
345 have a v-table). Because they are used so often, you must know what they
346 do and how they work. All of these templates are defined in the <a
347 href="/doxygen/Casting_8h-source.html"><tt>llvm/Support/Casting.h</tt></a>
348 file (note that you very rarely have to include this file directly).</p>
351 <dt><tt>isa<></tt>: </dt>
353 <dd><p>The <tt>isa<></tt> operator works exactly like the Java
354 "<tt>instanceof</tt>" operator. It returns true or false depending on whether
355 a reference or pointer points to an instance of the specified class. This can
356 be very useful for constraint checking of various sorts (example below).</p>
359 <dt><tt>cast<></tt>: </dt>
361 <dd><p>The <tt>cast<></tt> operator is a "checked cast" operation. It
362 converts a pointer or reference from a base class to a derived class, causing
363 an assertion failure if it is not really an instance of the right type. This
364 should be used in cases where you have some information that makes you believe
365 that something is of the right type. An example of the <tt>isa<></tt>
366 and <tt>cast<></tt> template is:</p>
368 <div class="doc_code">
370 static bool isLoopInvariant(const <a href="#Value">Value</a> *V, const Loop *L) {
371 if (isa<<a href="#Constant">Constant</a>>(V) || isa<<a href="#Argument">Argument</a>>(V) || isa<<a href="#GlobalValue">GlobalValue</a>>(V))
374 // <i>Otherwise, it must be an instruction...</i>
375 return !L->contains(cast<<a href="#Instruction">Instruction</a>>(V)->getParent());
380 <p>Note that you should <b>not</b> use an <tt>isa<></tt> test followed
381 by a <tt>cast<></tt>, for that use the <tt>dyn_cast<></tt>
386 <dt><tt>dyn_cast<></tt>:</dt>
388 <dd><p>The <tt>dyn_cast<></tt> operator is a "checking cast" operation.
389 It checks to see if the operand is of the specified type, and if so, returns a
390 pointer to it (this operator does not work with references). If the operand is
391 not of the correct type, a null pointer is returned. Thus, this works very
392 much like the <tt>dynamic_cast<></tt> operator in C++, and should be
393 used in the same circumstances. Typically, the <tt>dyn_cast<></tt>
394 operator is used in an <tt>if</tt> statement or some other flow control
395 statement like this:</p>
397 <div class="doc_code">
399 if (<a href="#AllocationInst">AllocationInst</a> *AI = dyn_cast<<a href="#AllocationInst">AllocationInst</a>>(Val)) {
405 <p>This form of the <tt>if</tt> statement effectively combines together a call
406 to <tt>isa<></tt> and a call to <tt>cast<></tt> into one
407 statement, which is very convenient.</p>
409 <p>Note that the <tt>dyn_cast<></tt> operator, like C++'s
410 <tt>dynamic_cast<></tt> or Java's <tt>instanceof</tt> operator, can be
411 abused. In particular, you should not use big chained <tt>if/then/else</tt>
412 blocks to check for lots of different variants of classes. If you find
413 yourself wanting to do this, it is much cleaner and more efficient to use the
414 <tt>InstVisitor</tt> class to dispatch over the instruction type directly.</p>
418 <dt><tt>cast_or_null<></tt>: </dt>
420 <dd><p>The <tt>cast_or_null<></tt> operator works just like the
421 <tt>cast<></tt> operator, except that it allows for a null pointer as an
422 argument (which it then propagates). This can sometimes be useful, allowing
423 you to combine several null checks into one.</p></dd>
425 <dt><tt>dyn_cast_or_null<></tt>: </dt>
427 <dd><p>The <tt>dyn_cast_or_null<></tt> operator works just like the
428 <tt>dyn_cast<></tt> operator, except that it allows for a null pointer
429 as an argument (which it then propagates). This can sometimes be useful,
430 allowing you to combine several null checks into one.</p></dd>
434 <p>These five templates can be used with any classes, whether they have a
435 v-table or not. To add support for these templates, you simply need to add
436 <tt>classof</tt> static methods to the class you are interested casting
437 to. Describing this is currently outside the scope of this document, but there
438 are lots of examples in the LLVM source base.</p>
443 <!-- ======================================================================= -->
445 <a name="string_apis">Passing strings (the <tt>StringRef</tt>
446 and <tt>Twine</tt> classes)</a>
451 <p>Although LLVM generally does not do much string manipulation, we do have
452 several important APIs which take strings. Two important examples are the
453 Value class -- which has names for instructions, functions, etc. -- and the
454 StringMap class which is used extensively in LLVM and Clang.</p>
456 <p>These are generic classes, and they need to be able to accept strings which
457 may have embedded null characters. Therefore, they cannot simply take
458 a <tt>const char *</tt>, and taking a <tt>const std::string&</tt> requires
459 clients to perform a heap allocation which is usually unnecessary. Instead,
460 many LLVM APIs use a <tt>StringRef</tt> or a <tt>const Twine&</tt> for
461 passing strings efficiently.</p>
463 <!-- _______________________________________________________________________ -->
465 <a name="StringRef">The <tt>StringRef</tt> class</a>
470 <p>The <tt>StringRef</tt> data type represents a reference to a constant string
471 (a character array and a length) and supports the common operations available
472 on <tt>std:string</tt>, but does not require heap allocation.</p>
474 <p>It can be implicitly constructed using a C style null-terminated string,
475 an <tt>std::string</tt>, or explicitly with a character pointer and length.
476 For example, the <tt>StringRef</tt> find function is declared as:</p>
478 <pre class="doc_code">
479 iterator find(StringRef Key);
482 <p>and clients can call it using any one of:</p>
484 <pre class="doc_code">
485 Map.find("foo"); <i>// Lookup "foo"</i>
486 Map.find(std::string("bar")); <i>// Lookup "bar"</i>
487 Map.find(StringRef("\0baz", 4)); <i>// Lookup "\0baz"</i>
490 <p>Similarly, APIs which need to return a string may return a <tt>StringRef</tt>
491 instance, which can be used directly or converted to an <tt>std::string</tt>
492 using the <tt>str</tt> member function. See
493 "<tt><a href="/doxygen/classllvm_1_1StringRef_8h-source.html">llvm/ADT/StringRef.h</a></tt>"
494 for more information.</p>
496 <p>You should rarely use the <tt>StringRef</tt> class directly, because it contains
497 pointers to external memory it is not generally safe to store an instance of the
498 class (unless you know that the external storage will not be freed). StringRef is
499 small and pervasive enough in LLVM that it should always be passed by value.</p>
503 <!-- _______________________________________________________________________ -->
505 <a name="Twine">The <tt>Twine</tt> class</a>
510 <p>The <tt>Twine</tt> class is an efficient way for APIs to accept concatenated
511 strings. For example, a common LLVM paradigm is to name one instruction based on
512 the name of another instruction with a suffix, for example:</p>
514 <div class="doc_code">
516 New = CmpInst::Create(<i>...</i>, SO->getName() + ".cmp");
520 <p>The <tt>Twine</tt> class is effectively a
521 lightweight <a href="http://en.wikipedia.org/wiki/Rope_(computer_science)">rope</a>
522 which points to temporary (stack allocated) objects. Twines can be implicitly
523 constructed as the result of the plus operator applied to strings (i.e., a C
524 strings, an <tt>std::string</tt>, or a <tt>StringRef</tt>). The twine delays the
525 actual concatenation of strings until it is actually required, at which point
526 it can be efficiently rendered directly into a character array. This avoids
527 unnecessary heap allocation involved in constructing the temporary results of
528 string concatenation. See
529 "<tt><a href="/doxygen/classllvm_1_1Twine_8h-source.html">llvm/ADT/Twine.h</a></tt>"
530 for more information.</p>
532 <p>As with a <tt>StringRef</tt>, <tt>Twine</tt> objects point to external memory
533 and should almost never be stored or mentioned directly. They are intended
534 solely for use when defining a function which should be able to efficiently
535 accept concatenated strings.</p>
541 <!-- ======================================================================= -->
543 <a name="DEBUG">The <tt>DEBUG()</tt> macro and <tt>-debug</tt> option</a>
548 <p>Often when working on your pass you will put a bunch of debugging printouts
549 and other code into your pass. After you get it working, you want to remove
550 it, but you may need it again in the future (to work out new bugs that you run
553 <p> Naturally, because of this, you don't want to delete the debug printouts,
554 but you don't want them to always be noisy. A standard compromise is to comment
555 them out, allowing you to enable them if you need them in the future.</p>
557 <p>The "<tt><a href="/doxygen/Debug_8h-source.html">llvm/Support/Debug.h</a></tt>"
558 file provides a macro named <tt>DEBUG()</tt> that is a much nicer solution to
559 this problem. Basically, you can put arbitrary code into the argument of the
560 <tt>DEBUG</tt> macro, and it is only executed if '<tt>opt</tt>' (or any other
561 tool) is run with the '<tt>-debug</tt>' command line argument:</p>
563 <div class="doc_code">
565 DEBUG(errs() << "I am here!\n");
569 <p>Then you can run your pass like this:</p>
571 <div class="doc_code">
573 $ opt < a.bc > /dev/null -mypass
574 <i><no output></i>
575 $ opt < a.bc > /dev/null -mypass -debug
580 <p>Using the <tt>DEBUG()</tt> macro instead of a home-brewed solution allows you
581 to not have to create "yet another" command line option for the debug output for
582 your pass. Note that <tt>DEBUG()</tt> macros are disabled for optimized builds,
583 so they do not cause a performance impact at all (for the same reason, they
584 should also not contain side-effects!).</p>
586 <p>One additional nice thing about the <tt>DEBUG()</tt> macro is that you can
587 enable or disable it directly in gdb. Just use "<tt>set DebugFlag=0</tt>" or
588 "<tt>set DebugFlag=1</tt>" from the gdb if the program is running. If the
589 program hasn't been started yet, you can always just run it with
592 <!-- _______________________________________________________________________ -->
594 <a name="DEBUG_TYPE">Fine grained debug info with <tt>DEBUG_TYPE</tt> and
595 the <tt>-debug-only</tt> option</a>
600 <p>Sometimes you may find yourself in a situation where enabling <tt>-debug</tt>
601 just turns on <b>too much</b> information (such as when working on the code
602 generator). If you want to enable debug information with more fine-grained
603 control, you define the <tt>DEBUG_TYPE</tt> macro and the <tt>-debug</tt> only
604 option as follows:</p>
606 <div class="doc_code">
609 DEBUG(errs() << "No debug type\n");
610 #define DEBUG_TYPE "foo"
611 DEBUG(errs() << "'foo' debug type\n");
613 #define DEBUG_TYPE "bar"
614 DEBUG(errs() << "'bar' debug type\n"));
616 #define DEBUG_TYPE ""
617 DEBUG(errs() << "No debug type (2)\n");
621 <p>Then you can run your pass like this:</p>
623 <div class="doc_code">
625 $ opt < a.bc > /dev/null -mypass
626 <i><no output></i>
627 $ opt < a.bc > /dev/null -mypass -debug
632 $ opt < a.bc > /dev/null -mypass -debug-only=foo
634 $ opt < a.bc > /dev/null -mypass -debug-only=bar
639 <p>Of course, in practice, you should only set <tt>DEBUG_TYPE</tt> at the top of
640 a file, to specify the debug type for the entire module (if you do this before
641 you <tt>#include "llvm/Support/Debug.h"</tt>, you don't have to insert the ugly
642 <tt>#undef</tt>'s). Also, you should use names more meaningful than "foo" and
643 "bar", because there is no system in place to ensure that names do not
644 conflict. If two different modules use the same string, they will all be turned
645 on when the name is specified. This allows, for example, all debug information
646 for instruction scheduling to be enabled with <tt>-debug-type=InstrSched</tt>,
647 even if the source lives in multiple files.</p>
649 <p>The <tt>DEBUG_WITH_TYPE</tt> macro is also available for situations where you
650 would like to set <tt>DEBUG_TYPE</tt>, but only for one specific <tt>DEBUG</tt>
651 statement. It takes an additional first parameter, which is the type to use. For
652 example, the preceding example could be written as:</p>
655 <div class="doc_code">
657 DEBUG_WITH_TYPE("", errs() << "No debug type\n");
658 DEBUG_WITH_TYPE("foo", errs() << "'foo' debug type\n");
659 DEBUG_WITH_TYPE("bar", errs() << "'bar' debug type\n"));
660 DEBUG_WITH_TYPE("", errs() << "No debug type (2)\n");
668 <!-- ======================================================================= -->
670 <a name="Statistic">The <tt>Statistic</tt> class & <tt>-stats</tt>
677 href="/doxygen/Statistic_8h-source.html">llvm/ADT/Statistic.h</a></tt>" file
678 provides a class named <tt>Statistic</tt> that is used as a unified way to
679 keep track of what the LLVM compiler is doing and how effective various
680 optimizations are. It is useful to see what optimizations are contributing to
681 making a particular program run faster.</p>
683 <p>Often you may run your pass on some big program, and you're interested to see
684 how many times it makes a certain transformation. Although you can do this with
685 hand inspection, or some ad-hoc method, this is a real pain and not very useful
686 for big programs. Using the <tt>Statistic</tt> class makes it very easy to
687 keep track of this information, and the calculated information is presented in a
688 uniform manner with the rest of the passes being executed.</p>
690 <p>There are many examples of <tt>Statistic</tt> uses, but the basics of using
691 it are as follows:</p>
694 <li><p>Define your statistic like this:</p>
696 <div class="doc_code">
698 #define <a href="#DEBUG_TYPE">DEBUG_TYPE</a> "mypassname" <i>// This goes before any #includes.</i>
699 STATISTIC(NumXForms, "The # of times I did stuff");
703 <p>The <tt>STATISTIC</tt> macro defines a static variable, whose name is
704 specified by the first argument. The pass name is taken from the DEBUG_TYPE
705 macro, and the description is taken from the second argument. The variable
706 defined ("NumXForms" in this case) acts like an unsigned integer.</p></li>
708 <li><p>Whenever you make a transformation, bump the counter:</p>
710 <div class="doc_code">
712 ++NumXForms; // <i>I did stuff!</i>
719 <p>That's all you have to do. To get '<tt>opt</tt>' to print out the
720 statistics gathered, use the '<tt>-stats</tt>' option:</p>
722 <div class="doc_code">
724 $ opt -stats -mypassname < program.bc > /dev/null
725 <i>... statistics output ...</i>
729 <p> When running <tt>opt</tt> on a C file from the SPEC benchmark
730 suite, it gives a report that looks like this:</p>
732 <div class="doc_code">
734 7646 bitcodewriter - Number of normal instructions
735 725 bitcodewriter - Number of oversized instructions
736 129996 bitcodewriter - Number of bitcode bytes written
737 2817 raise - Number of insts DCEd or constprop'd
738 3213 raise - Number of cast-of-self removed
739 5046 raise - Number of expression trees converted
740 75 raise - Number of other getelementptr's formed
741 138 raise - Number of load/store peepholes
742 42 deadtypeelim - Number of unused typenames removed from symtab
743 392 funcresolve - Number of varargs functions resolved
744 27 globaldce - Number of global variables removed
745 2 adce - Number of basic blocks removed
746 134 cee - Number of branches revectored
747 49 cee - Number of setcc instruction eliminated
748 532 gcse - Number of loads removed
749 2919 gcse - Number of instructions removed
750 86 indvars - Number of canonical indvars added
751 87 indvars - Number of aux indvars removed
752 25 instcombine - Number of dead inst eliminate
753 434 instcombine - Number of insts combined
754 248 licm - Number of load insts hoisted
755 1298 licm - Number of insts hoisted to a loop pre-header
756 3 licm - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
757 75 mem2reg - Number of alloca's promoted
758 1444 cfgsimplify - Number of blocks simplified
762 <p>Obviously, with so many optimizations, having a unified framework for this
763 stuff is very nice. Making your pass fit well into the framework makes it more
764 maintainable and useful.</p>
768 <!-- ======================================================================= -->
770 <a name="ViewGraph">Viewing graphs while debugging code</a>
775 <p>Several of the important data structures in LLVM are graphs: for example
776 CFGs made out of LLVM <a href="#BasicBlock">BasicBlock</a>s, CFGs made out of
777 LLVM <a href="CodeGenerator.html#machinebasicblock">MachineBasicBlock</a>s, and
778 <a href="CodeGenerator.html#selectiondag_intro">Instruction Selection
779 DAGs</a>. In many cases, while debugging various parts of the compiler, it is
780 nice to instantly visualize these graphs.</p>
782 <p>LLVM provides several callbacks that are available in a debug build to do
783 exactly that. If you call the <tt>Function::viewCFG()</tt> method, for example,
784 the current LLVM tool will pop up a window containing the CFG for the function
785 where each basic block is a node in the graph, and each node contains the
786 instructions in the block. Similarly, there also exists
787 <tt>Function::viewCFGOnly()</tt> (does not include the instructions), the
788 <tt>MachineFunction::viewCFG()</tt> and <tt>MachineFunction::viewCFGOnly()</tt>,
789 and the <tt>SelectionDAG::viewGraph()</tt> methods. Within GDB, for example,
790 you can usually use something like <tt>call DAG.viewGraph()</tt> to pop
791 up a window. Alternatively, you can sprinkle calls to these functions in your
792 code in places you want to debug.</p>
794 <p>Getting this to work requires a small amount of configuration. On Unix
795 systems with X11, install the <a href="http://www.graphviz.org">graphviz</a>
796 toolkit, and make sure 'dot' and 'gv' are in your path. If you are running on
797 Mac OS/X, download and install the Mac OS/X <a
798 href="http://www.pixelglow.com/graphviz/">Graphviz program</a>, and add
799 <tt>/Applications/Graphviz.app/Contents/MacOS/</tt> (or wherever you install
800 it) to your path. Once in your system and path are set up, rerun the LLVM
801 configure script and rebuild LLVM to enable this functionality.</p>
803 <p><tt>SelectionDAG</tt> has been extended to make it easier to locate
804 <i>interesting</i> nodes in large complex graphs. From gdb, if you
805 <tt>call DAG.setGraphColor(<i>node</i>, "<i>color</i>")</tt>, then the
806 next <tt>call DAG.viewGraph()</tt> would highlight the node in the
807 specified color (choices of colors can be found at <a
808 href="http://www.graphviz.org/doc/info/colors.html">colors</a>.) More
809 complex node attributes can be provided with <tt>call
810 DAG.setGraphAttrs(<i>node</i>, "<i>attributes</i>")</tt> (choices can be
811 found at <a href="http://www.graphviz.org/doc/info/attrs.html">Graph
812 Attributes</a>.) If you want to restart and clear all the current graph
813 attributes, then you can <tt>call DAG.clearGraphAttrs()</tt>. </p>
815 <p>Note that graph visualization features are compiled out of Release builds
816 to reduce file size. This means that you need a Debug+Asserts or
817 Release+Asserts build to use these features.</p>
823 <!-- *********************************************************************** -->
825 <a name="datastructure">Picking the Right Data Structure for a Task</a>
827 <!-- *********************************************************************** -->
831 <p>LLVM has a plethora of data structures in the <tt>llvm/ADT/</tt> directory,
832 and we commonly use STL data structures. This section describes the trade-offs
833 you should consider when you pick one.</p>
836 The first step is a choose your own adventure: do you want a sequential
837 container, a set-like container, or a map-like container? The most important
838 thing when choosing a container is the algorithmic properties of how you plan to
839 access the container. Based on that, you should use:</p>
842 <li>a <a href="#ds_map">map-like</a> container if you need efficient look-up
843 of an value based on another value. Map-like containers also support
844 efficient queries for containment (whether a key is in the map). Map-like
845 containers generally do not support efficient reverse mapping (values to
846 keys). If you need that, use two maps. Some map-like containers also
847 support efficient iteration through the keys in sorted order. Map-like
848 containers are the most expensive sort, only use them if you need one of
849 these capabilities.</li>
851 <li>a <a href="#ds_set">set-like</a> container if you need to put a bunch of
852 stuff into a container that automatically eliminates duplicates. Some
853 set-like containers support efficient iteration through the elements in
854 sorted order. Set-like containers are more expensive than sequential
858 <li>a <a href="#ds_sequential">sequential</a> container provides
859 the most efficient way to add elements and keeps track of the order they are
860 added to the collection. They permit duplicates and support efficient
861 iteration, but do not support efficient look-up based on a key.
864 <li>a <a href="#ds_string">string</a> container is a specialized sequential
865 container or reference structure that is used for character or byte
868 <li>a <a href="#ds_bit">bit</a> container provides an efficient way to store and
869 perform set operations on sets of numeric id's, while automatically
870 eliminating duplicates. Bit containers require a maximum of 1 bit for each
871 identifier you want to store.
876 Once the proper category of container is determined, you can fine tune the
877 memory use, constant factors, and cache behaviors of access by intelligently
878 picking a member of the category. Note that constant factors and cache behavior
879 can be a big deal. If you have a vector that usually only contains a few
880 elements (but could contain many), for example, it's much better to use
881 <a href="#dss_smallvector">SmallVector</a> than <a href="#dss_vector">vector</a>
882 . Doing so avoids (relatively) expensive malloc/free calls, which dwarf the
883 cost of adding the elements to the container. </p>
885 <!-- ======================================================================= -->
887 <a name="ds_sequential">Sequential Containers (std::vector, std::list, etc)</a>
891 There are a variety of sequential containers available for you, based on your
892 needs. Pick the first in this section that will do what you want.
894 <!-- _______________________________________________________________________ -->
896 <a name="dss_arrayref">llvm/ADT/ArrayRef.h</a>
900 <p>The llvm::ArrayRef class is the preferred class to use in an interface that
901 accepts a sequential list of elements in memory and just reads from them. By
902 taking an ArrayRef, the API can be passed a fixed size array, an std::vector,
903 an llvm::SmallVector and anything else that is contiguous in memory.
909 <!-- _______________________________________________________________________ -->
911 <a name="dss_fixedarrays">Fixed Size Arrays</a>
915 <p>Fixed size arrays are very simple and very fast. They are good if you know
916 exactly how many elements you have, or you have a (low) upper bound on how many
920 <!-- _______________________________________________________________________ -->
922 <a name="dss_heaparrays">Heap Allocated Arrays</a>
926 <p>Heap allocated arrays (new[] + delete[]) are also simple. They are good if
927 the number of elements is variable, if you know how many elements you will need
928 before the array is allocated, and if the array is usually large (if not,
929 consider a <a href="#dss_smallvector">SmallVector</a>). The cost of a heap
930 allocated array is the cost of the new/delete (aka malloc/free). Also note that
931 if you are allocating an array of a type with a constructor, the constructor and
932 destructors will be run for every element in the array (re-sizable vectors only
933 construct those elements actually used).</p>
936 <!-- _______________________________________________________________________ -->
938 <a name="dss_tinyptrvector">"llvm/ADT/TinyPtrVector.h"</a>
943 <p><tt>TinyPtrVector<Type></tt> is a highly specialized collection class
944 that is optimized to avoid allocation in the case when a vector has zero or one
945 elements. It has two major restrictions: 1) it can only hold values of pointer
946 type, and 2) it cannot hold a null pointer.</p>
948 <p>Since this container is highly specialized, it is rarely used.</p>
952 <!-- _______________________________________________________________________ -->
954 <a name="dss_smallvector">"llvm/ADT/SmallVector.h"</a>
958 <p><tt>SmallVector<Type, N></tt> is a simple class that looks and smells
959 just like <tt>vector<Type></tt>:
960 it supports efficient iteration, lays out elements in memory order (so you can
961 do pointer arithmetic between elements), supports efficient push_back/pop_back
962 operations, supports efficient random access to its elements, etc.</p>
964 <p>The advantage of SmallVector is that it allocates space for
965 some number of elements (N) <b>in the object itself</b>. Because of this, if
966 the SmallVector is dynamically smaller than N, no malloc is performed. This can
967 be a big win in cases where the malloc/free call is far more expensive than the
968 code that fiddles around with the elements.</p>
970 <p>This is good for vectors that are "usually small" (e.g. the number of
971 predecessors/successors of a block is usually less than 8). On the other hand,
972 this makes the size of the SmallVector itself large, so you don't want to
973 allocate lots of them (doing so will waste a lot of space). As such,
974 SmallVectors are most useful when on the stack.</p>
976 <p>SmallVector also provides a nice portable and efficient replacement for
981 <!-- _______________________________________________________________________ -->
983 <a name="dss_vector"><vector></a>
988 std::vector is well loved and respected. It is useful when SmallVector isn't:
989 when the size of the vector is often large (thus the small optimization will
990 rarely be a benefit) or if you will be allocating many instances of the vector
991 itself (which would waste space for elements that aren't in the container).
992 vector is also useful when interfacing with code that expects vectors :).
995 <p>One worthwhile note about std::vector: avoid code like this:</p>
997 <div class="doc_code">
1000 std::vector<foo> V;
1006 <p>Instead, write this as:</p>
1008 <div class="doc_code">
1010 std::vector<foo> V;
1018 <p>Doing so will save (at least) one heap allocation and free per iteration of
1023 <!-- _______________________________________________________________________ -->
1025 <a name="dss_deque"><deque></a>
1029 <p>std::deque is, in some senses, a generalized version of std::vector. Like
1030 std::vector, it provides constant time random access and other similar
1031 properties, but it also provides efficient access to the front of the list. It
1032 does not guarantee continuity of elements within memory.</p>
1034 <p>In exchange for this extra flexibility, std::deque has significantly higher
1035 constant factor costs than std::vector. If possible, use std::vector or
1036 something cheaper.</p>
1039 <!-- _______________________________________________________________________ -->
1041 <a name="dss_list"><list></a>
1045 <p>std::list is an extremely inefficient class that is rarely useful.
1046 It performs a heap allocation for every element inserted into it, thus having an
1047 extremely high constant factor, particularly for small data types. std::list
1048 also only supports bidirectional iteration, not random access iteration.</p>
1050 <p>In exchange for this high cost, std::list supports efficient access to both
1051 ends of the list (like std::deque, but unlike std::vector or SmallVector). In
1052 addition, the iterator invalidation characteristics of std::list are stronger
1053 than that of a vector class: inserting or removing an element into the list does
1054 not invalidate iterator or pointers to other elements in the list.</p>
1057 <!-- _______________________________________________________________________ -->
1059 <a name="dss_ilist">llvm/ADT/ilist.h</a>
1063 <p><tt>ilist<T></tt> implements an 'intrusive' doubly-linked list. It is
1064 intrusive, because it requires the element to store and provide access to the
1065 prev/next pointers for the list.</p>
1067 <p><tt>ilist</tt> has the same drawbacks as <tt>std::list</tt>, and additionally
1068 requires an <tt>ilist_traits</tt> implementation for the element type, but it
1069 provides some novel characteristics. In particular, it can efficiently store
1070 polymorphic objects, the traits class is informed when an element is inserted or
1071 removed from the list, and <tt>ilist</tt>s are guaranteed to support a
1072 constant-time splice operation.</p>
1074 <p>These properties are exactly what we want for things like
1075 <tt>Instruction</tt>s and basic blocks, which is why these are implemented with
1076 <tt>ilist</tt>s.</p>
1078 Related classes of interest are explained in the following subsections:
1080 <li><a href="#dss_ilist_traits">ilist_traits</a></li>
1081 <li><a href="#dss_iplist">iplist</a></li>
1082 <li><a href="#dss_ilist_node">llvm/ADT/ilist_node.h</a></li>
1083 <li><a href="#dss_ilist_sentinel">Sentinels</a></li>
1087 <!-- _______________________________________________________________________ -->
1089 <a name="dss_packedvector">llvm/ADT/PackedVector.h</a>
1094 Useful for storing a vector of values using only a few number of bits for each
1095 value. Apart from the standard operations of a vector-like container, it can
1096 also perform an 'or' set operation.
1101 <div class="doc_code">
1105 FirstCondition = 0x1,
1106 SecondCondition = 0x2,
1111 PackedVector<State, 2> Vec1;
1112 Vec1.push_back(FirstCondition);
1114 PackedVector<State, 2> Vec2;
1115 Vec2.push_back(SecondCondition);
1118 return Vec1[0]; // returns 'Both'.
1125 <!-- _______________________________________________________________________ -->
1127 <a name="dss_ilist_traits">ilist_traits</a>
1131 <p><tt>ilist_traits<T></tt> is <tt>ilist<T></tt>'s customization
1132 mechanism. <tt>iplist<T></tt> (and consequently <tt>ilist<T></tt>)
1133 publicly derive from this traits class.</p>
1136 <!-- _______________________________________________________________________ -->
1138 <a name="dss_iplist">iplist</a>
1142 <p><tt>iplist<T></tt> is <tt>ilist<T></tt>'s base and as such
1143 supports a slightly narrower interface. Notably, inserters from
1144 <tt>T&</tt> are absent.</p>
1146 <p><tt>ilist_traits<T></tt> is a public base of this class and can be
1147 used for a wide variety of customizations.</p>
1150 <!-- _______________________________________________________________________ -->
1152 <a name="dss_ilist_node">llvm/ADT/ilist_node.h</a>
1156 <p><tt>ilist_node<T></tt> implements a the forward and backward links
1157 that are expected by the <tt>ilist<T></tt> (and analogous containers)
1158 in the default manner.</p>
1160 <p><tt>ilist_node<T></tt>s are meant to be embedded in the node type
1161 <tt>T</tt>, usually <tt>T</tt> publicly derives from
1162 <tt>ilist_node<T></tt>.</p>
1165 <!-- _______________________________________________________________________ -->
1167 <a name="dss_ilist_sentinel">Sentinels</a>
1171 <p><tt>ilist</tt>s have another specialty that must be considered. To be a good
1172 citizen in the C++ ecosystem, it needs to support the standard container
1173 operations, such as <tt>begin</tt> and <tt>end</tt> iterators, etc. Also, the
1174 <tt>operator--</tt> must work correctly on the <tt>end</tt> iterator in the
1175 case of non-empty <tt>ilist</tt>s.</p>
1177 <p>The only sensible solution to this problem is to allocate a so-called
1178 <i>sentinel</i> along with the intrusive list, which serves as the <tt>end</tt>
1179 iterator, providing the back-link to the last element. However conforming to the
1180 C++ convention it is illegal to <tt>operator++</tt> beyond the sentinel and it
1181 also must not be dereferenced.</p>
1183 <p>These constraints allow for some implementation freedom to the <tt>ilist</tt>
1184 how to allocate and store the sentinel. The corresponding policy is dictated
1185 by <tt>ilist_traits<T></tt>. By default a <tt>T</tt> gets heap-allocated
1186 whenever the need for a sentinel arises.</p>
1188 <p>While the default policy is sufficient in most cases, it may break down when
1189 <tt>T</tt> does not provide a default constructor. Also, in the case of many
1190 instances of <tt>ilist</tt>s, the memory overhead of the associated sentinels
1191 is wasted. To alleviate the situation with numerous and voluminous
1192 <tt>T</tt>-sentinels, sometimes a trick is employed, leading to <i>ghostly
1195 <p>Ghostly sentinels are obtained by specially-crafted <tt>ilist_traits<T></tt>
1196 which superpose the sentinel with the <tt>ilist</tt> instance in memory. Pointer
1197 arithmetic is used to obtain the sentinel, which is relative to the
1198 <tt>ilist</tt>'s <tt>this</tt> pointer. The <tt>ilist</tt> is augmented by an
1199 extra pointer, which serves as the back-link of the sentinel. This is the only
1200 field in the ghostly sentinel which can be legally accessed.</p>
1203 <!-- _______________________________________________________________________ -->
1205 <a name="dss_other">Other Sequential Container options</a>
1209 <p>Other STL containers are available, such as std::string.</p>
1211 <p>There are also various STL adapter classes such as std::queue,
1212 std::priority_queue, std::stack, etc. These provide simplified access to an
1213 underlying container but don't affect the cost of the container itself.</p>
1218 <!-- ======================================================================= -->
1220 <a name="ds_string">String-like containers</a>
1226 There are a variety of ways to pass around and use strings in C and C++, and
1227 LLVM adds a few new options to choose from. Pick the first option on this list
1228 that will do what you need, they are ordered according to their relative cost.
1231 Note that is is generally preferred to <em>not</em> pass strings around as
1232 "<tt>const char*</tt>"'s. These have a number of problems, including the fact
1233 that they cannot represent embedded nul ("\0") characters, and do not have a
1234 length available efficiently. The general replacement for '<tt>const
1235 char*</tt>' is StringRef.
1238 <p>For more information on choosing string containers for APIs, please see
1239 <a href="#string_apis">Passing strings</a>.</p>
1242 <!-- _______________________________________________________________________ -->
1244 <a name="dss_stringref">llvm/ADT/StringRef.h</a>
1249 The StringRef class is a simple value class that contains a pointer to a
1250 character and a length, and is quite related to the <a
1251 href="#dss_arrayref">ArrayRef</a> class (but specialized for arrays of
1252 characters). Because StringRef carries a length with it, it safely handles
1253 strings with embedded nul characters in it, getting the length does not require
1254 a strlen call, and it even has very convenient APIs for slicing and dicing the
1255 character range that it represents.
1259 StringRef is ideal for passing simple strings around that are known to be live,
1260 either because they are C string literals, std::string, a C array, or a
1261 SmallVector. Each of these cases has an efficient implicit conversion to
1262 StringRef, which doesn't result in a dynamic strlen being executed.
1265 <p>StringRef has a few major limitations which make more powerful string
1266 containers useful:</p>
1269 <li>You cannot directly convert a StringRef to a 'const char*' because there is
1270 no way to add a trailing nul (unlike the .c_str() method on various stronger
1274 <li>StringRef doesn't own or keep alive the underlying string bytes.
1275 As such it can easily lead to dangling pointers, and is not suitable for
1276 embedding in datastructures in most cases (instead, use an std::string or
1277 something like that).</li>
1279 <li>For the same reason, StringRef cannot be used as the return value of a
1280 method if the method "computes" the result string. Instead, use
1283 <li>StringRef's do not allow you to mutate the pointed-to string bytes and it
1284 doesn't allow you to insert or remove bytes from the range. For editing
1285 operations like this, it interoperates with the <a
1286 href="#dss_twine">Twine</a> class.</li>
1289 <p>Because of its strengths and limitations, it is very common for a function to
1290 take a StringRef and for a method on an object to return a StringRef that
1291 points into some string that it owns.</p>
1295 <!-- _______________________________________________________________________ -->
1297 <a name="dss_twine">llvm/ADT/Twine.h</a>
1302 The Twine class is used as an intermediary datatype for APIs that want to take
1303 a string that can be constructed inline with a series of concatenations.
1304 Twine works by forming recursive instances of the Twine datatype (a simple
1305 value object) on the stack as temporary objects, linking them together into a
1306 tree which is then linearized when the Twine is consumed. Twine is only safe
1307 to use as the argument to a function, and should always be a const reference,
1312 void foo(const Twine &T);
1316 foo(X + "." + Twine(i));
1319 <p>This example forms a string like "blarg.42" by concatenating the values
1320 together, and does not form intermediate strings containing "blarg" or
1324 <p>Because Twine is constructed with temporary objects on the stack, and
1325 because these instances are destroyed at the end of the current statement,
1326 it is an inherently dangerous API. For example, this simple variant contains
1327 undefined behavior and will probably crash:</p>
1330 void foo(const Twine &T);
1334 const Twine &Tmp = X + "." + Twine(i);
1338 <p>... because the temporaries are destroyed before the call. That said,
1339 Twine's are much more efficient than intermediate std::string temporaries, and
1340 they work really well with StringRef. Just be aware of their limitations.</p>
1345 <!-- _______________________________________________________________________ -->
1347 <a name="dss_smallstring">llvm/ADT/SmallString.h</a>
1352 <p>SmallString is a subclass of <a href="#dss_smallvector">SmallVector</a> that
1353 adds some convenience APIs like += that takes StringRef's. SmallString avoids
1354 allocating memory in the case when the preallocated space is enough to hold its
1355 data, and it calls back to general heap allocation when required. Since it owns
1356 its data, it is very safe to use and supports full mutation of the string.</p>
1358 <p>Like SmallVector's, the big downside to SmallString is their sizeof. While
1359 they are optimized for small strings, they themselves are not particularly
1360 small. This means that they work great for temporary scratch buffers on the
1361 stack, but should not generally be put into the heap: it is very rare to
1362 see a SmallString as the member of a frequently-allocated heap data structure
1363 or returned by-value.
1368 <!-- _______________________________________________________________________ -->
1370 <a name="dss_stdstring">std::string</a>
1375 <p>The standard C++ std::string class is a very general class that (like
1376 SmallString) owns its underlying data. sizeof(std::string) is very reasonable
1377 so it can be embedded into heap data structures and returned by-value.
1378 On the other hand, std::string is highly inefficient for inline editing (e.g.
1379 concatenating a bunch of stuff together) and because it is provided by the
1380 standard library, its performance characteristics depend a lot of the host
1381 standard library (e.g. libc++ and MSVC provide a highly optimized string
1382 class, GCC contains a really slow implementation).
1385 <p>The major disadvantage of std::string is that almost every operation that
1386 makes them larger can allocate memory, which is slow. As such, it is better
1387 to use SmallVector or Twine as a scratch buffer, but then use std::string to
1388 persist the result.</p>
1393 <!-- end of strings -->
1397 <!-- ======================================================================= -->
1399 <a name="ds_set">Set-Like Containers (std::set, SmallSet, SetVector, etc)</a>
1404 <p>Set-like containers are useful when you need to canonicalize multiple values
1405 into a single representation. There are several different choices for how to do
1406 this, providing various trade-offs.</p>
1408 <!-- _______________________________________________________________________ -->
1410 <a name="dss_sortedvectorset">A sorted 'vector'</a>
1415 <p>If you intend to insert a lot of elements, then do a lot of queries, a
1416 great approach is to use a vector (or other sequential container) with
1417 std::sort+std::unique to remove duplicates. This approach works really well if
1418 your usage pattern has these two distinct phases (insert then query), and can be
1419 coupled with a good choice of <a href="#ds_sequential">sequential container</a>.
1423 This combination provides the several nice properties: the result data is
1424 contiguous in memory (good for cache locality), has few allocations, is easy to
1425 address (iterators in the final vector are just indices or pointers), and can be
1426 efficiently queried with a standard binary or radix search.</p>
1430 <!-- _______________________________________________________________________ -->
1432 <a name="dss_smallset">"llvm/ADT/SmallSet.h"</a>
1437 <p>If you have a set-like data structure that is usually small and whose elements
1438 are reasonably small, a <tt>SmallSet<Type, N></tt> is a good choice. This set
1439 has space for N elements in place (thus, if the set is dynamically smaller than
1440 N, no malloc traffic is required) and accesses them with a simple linear search.
1441 When the set grows beyond 'N' elements, it allocates a more expensive representation that
1442 guarantees efficient access (for most types, it falls back to std::set, but for
1443 pointers it uses something far better, <a
1444 href="#dss_smallptrset">SmallPtrSet</a>).</p>
1446 <p>The magic of this class is that it handles small sets extremely efficiently,
1447 but gracefully handles extremely large sets without loss of efficiency. The
1448 drawback is that the interface is quite small: it supports insertion, queries
1449 and erasing, but does not support iteration.</p>
1453 <!-- _______________________________________________________________________ -->
1455 <a name="dss_smallptrset">"llvm/ADT/SmallPtrSet.h"</a>
1460 <p>SmallPtrSet has all the advantages of <tt>SmallSet</tt> (and a <tt>SmallSet</tt> of pointers is
1461 transparently implemented with a <tt>SmallPtrSet</tt>), but also supports iterators. If
1462 more than 'N' insertions are performed, a single quadratically
1463 probed hash table is allocated and grows as needed, providing extremely
1464 efficient access (constant time insertion/deleting/queries with low constant
1465 factors) and is very stingy with malloc traffic.</p>
1467 <p>Note that, unlike <tt>std::set</tt>, the iterators of <tt>SmallPtrSet</tt> are invalidated
1468 whenever an insertion occurs. Also, the values visited by the iterators are not
1469 visited in sorted order.</p>
1473 <!-- _______________________________________________________________________ -->
1475 <a name="dss_denseset">"llvm/ADT/DenseSet.h"</a>
1481 DenseSet is a simple quadratically probed hash table. It excels at supporting
1482 small values: it uses a single allocation to hold all of the pairs that
1483 are currently inserted in the set. DenseSet is a great way to unique small
1484 values that are not simple pointers (use <a
1485 href="#dss_smallptrset">SmallPtrSet</a> for pointers). Note that DenseSet has
1486 the same requirements for the value type that <a
1487 href="#dss_densemap">DenseMap</a> has.
1492 <!-- _______________________________________________________________________ -->
1494 <a name="dss_sparseset">"llvm/ADT/SparseSet.h"</a>
1499 <p>SparseSet holds a small number of objects identified by unsigned keys of
1500 moderate size. It uses a lot of memory, but provides operations that are
1501 almost as fast as a vector. Typical keys are physical registers, virtual
1502 registers, or numbered basic blocks.</p>
1504 <p>SparseSet is useful for algorithms that need very fast clear/find/insert/erase
1505 and fast iteration over small sets. It is not intended for building composite
1506 data structures.</p>
1510 <!-- _______________________________________________________________________ -->
1512 <a name="dss_FoldingSet">"llvm/ADT/FoldingSet.h"</a>
1518 FoldingSet is an aggregate class that is really good at uniquing
1519 expensive-to-create or polymorphic objects. It is a combination of a chained
1520 hash table with intrusive links (uniqued objects are required to inherit from
1521 FoldingSetNode) that uses <a href="#dss_smallvector">SmallVector</a> as part of
1524 <p>Consider a case where you want to implement a "getOrCreateFoo" method for
1525 a complex object (for example, a node in the code generator). The client has a
1526 description of *what* it wants to generate (it knows the opcode and all the
1527 operands), but we don't want to 'new' a node, then try inserting it into a set
1528 only to find out it already exists, at which point we would have to delete it
1529 and return the node that already exists.
1532 <p>To support this style of client, FoldingSet perform a query with a
1533 FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
1534 element that we want to query for. The query either returns the element
1535 matching the ID or it returns an opaque ID that indicates where insertion should
1536 take place. Construction of the ID usually does not require heap traffic.</p>
1538 <p>Because FoldingSet uses intrusive links, it can support polymorphic objects
1539 in the set (for example, you can have SDNode instances mixed with LoadSDNodes).
1540 Because the elements are individually allocated, pointers to the elements are
1541 stable: inserting or removing elements does not invalidate any pointers to other
1547 <!-- _______________________________________________________________________ -->
1549 <a name="dss_set"><set></a>
1554 <p><tt>std::set</tt> is a reasonable all-around set class, which is decent at
1555 many things but great at nothing. std::set allocates memory for each element
1556 inserted (thus it is very malloc intensive) and typically stores three pointers
1557 per element in the set (thus adding a large amount of per-element space
1558 overhead). It offers guaranteed log(n) performance, which is not particularly
1559 fast from a complexity standpoint (particularly if the elements of the set are
1560 expensive to compare, like strings), and has extremely high constant factors for
1561 lookup, insertion and removal.</p>
1563 <p>The advantages of std::set are that its iterators are stable (deleting or
1564 inserting an element from the set does not affect iterators or pointers to other
1565 elements) and that iteration over the set is guaranteed to be in sorted order.
1566 If the elements in the set are large, then the relative overhead of the pointers
1567 and malloc traffic is not a big deal, but if the elements of the set are small,
1568 std::set is almost never a good choice.</p>
1572 <!-- _______________________________________________________________________ -->
1574 <a name="dss_setvector">"llvm/ADT/SetVector.h"</a>
1578 <p>LLVM's SetVector<Type> is an adapter class that combines your choice of
1579 a set-like container along with a <a href="#ds_sequential">Sequential
1580 Container</a>. The important property
1581 that this provides is efficient insertion with uniquing (duplicate elements are
1582 ignored) with iteration support. It implements this by inserting elements into
1583 both a set-like container and the sequential container, using the set-like
1584 container for uniquing and the sequential container for iteration.
1587 <p>The difference between SetVector and other sets is that the order of
1588 iteration is guaranteed to match the order of insertion into the SetVector.
1589 This property is really important for things like sets of pointers. Because
1590 pointer values are non-deterministic (e.g. vary across runs of the program on
1591 different machines), iterating over the pointers in the set will
1592 not be in a well-defined order.</p>
1595 The drawback of SetVector is that it requires twice as much space as a normal
1596 set and has the sum of constant factors from the set-like container and the
1597 sequential container that it uses. Use it *only* if you need to iterate over
1598 the elements in a deterministic order. SetVector is also expensive to delete
1599 elements out of (linear time), unless you use it's "pop_back" method, which is
1603 <p><tt>SetVector</tt> is an adapter class that defaults to
1604 using <tt>std::vector</tt> and a size 16 <tt>SmallSet</tt> for the underlying
1605 containers, so it is quite expensive. However,
1606 <tt>"llvm/ADT/SetVector.h"</tt> also provides a <tt>SmallSetVector</tt>
1607 class, which defaults to using a <tt>SmallVector</tt> and <tt>SmallSet</tt>
1608 of a specified size. If you use this, and if your sets are dynamically
1609 smaller than <tt>N</tt>, you will save a lot of heap traffic.</p>
1613 <!-- _______________________________________________________________________ -->
1615 <a name="dss_uniquevector">"llvm/ADT/UniqueVector.h"</a>
1621 UniqueVector is similar to <a href="#dss_setvector">SetVector</a>, but it
1622 retains a unique ID for each element inserted into the set. It internally
1623 contains a map and a vector, and it assigns a unique ID for each value inserted
1626 <p>UniqueVector is very expensive: its cost is the sum of the cost of
1627 maintaining both the map and vector, it has high complexity, high constant
1628 factors, and produces a lot of malloc traffic. It should be avoided.</p>
1632 <!-- _______________________________________________________________________ -->
1634 <a name="dss_immutableset">"llvm/ADT/ImmutableSet.h"</a>
1640 ImmutableSet is an immutable (functional) set implementation based on an AVL
1642 Adding or removing elements is done through a Factory object and results in the
1643 creation of a new ImmutableSet object.
1644 If an ImmutableSet already exists with the given contents, then the existing one
1645 is returned; equality is compared with a FoldingSetNodeID.
1646 The time and space complexity of add or remove operations is logarithmic in the
1647 size of the original set.
1650 There is no method for returning an element of the set, you can only check for
1656 <!-- _______________________________________________________________________ -->
1658 <a name="dss_otherset">Other Set-Like Container Options</a>
1664 The STL provides several other options, such as std::multiset and the various
1665 "hash_set" like containers (whether from C++ TR1 or from the SGI library). We
1666 never use hash_set and unordered_set because they are generally very expensive
1667 (each insertion requires a malloc) and very non-portable.
1670 <p>std::multiset is useful if you're not interested in elimination of
1671 duplicates, but has all the drawbacks of std::set. A sorted vector (where you
1672 don't delete duplicate entries) or some other approach is almost always
1679 <!-- ======================================================================= -->
1681 <a name="ds_map">Map-Like Containers (std::map, DenseMap, etc)</a>
1685 Map-like containers are useful when you want to associate data to a key. As
1686 usual, there are a lot of different ways to do this. :)
1688 <!-- _______________________________________________________________________ -->
1690 <a name="dss_sortedvectormap">A sorted 'vector'</a>
1696 If your usage pattern follows a strict insert-then-query approach, you can
1697 trivially use the same approach as <a href="#dss_sortedvectorset">sorted vectors
1698 for set-like containers</a>. The only difference is that your query function
1699 (which uses std::lower_bound to get efficient log(n) lookup) should only compare
1700 the key, not both the key and value. This yields the same advantages as sorted
1705 <!-- _______________________________________________________________________ -->
1707 <a name="dss_stringmap">"llvm/ADT/StringMap.h"</a>
1713 Strings are commonly used as keys in maps, and they are difficult to support
1714 efficiently: they are variable length, inefficient to hash and compare when
1715 long, expensive to copy, etc. StringMap is a specialized container designed to
1716 cope with these issues. It supports mapping an arbitrary range of bytes to an
1717 arbitrary other object.</p>
1719 <p>The StringMap implementation uses a quadratically-probed hash table, where
1720 the buckets store a pointer to the heap allocated entries (and some other
1721 stuff). The entries in the map must be heap allocated because the strings are
1722 variable length. The string data (key) and the element object (value) are
1723 stored in the same allocation with the string data immediately after the element
1724 object. This container guarantees the "<tt>(char*)(&Value+1)</tt>" points
1725 to the key string for a value.</p>
1727 <p>The StringMap is very fast for several reasons: quadratic probing is very
1728 cache efficient for lookups, the hash value of strings in buckets is not
1729 recomputed when looking up an element, StringMap rarely has to touch the
1730 memory for unrelated objects when looking up a value (even when hash collisions
1731 happen), hash table growth does not recompute the hash values for strings
1732 already in the table, and each pair in the map is store in a single allocation
1733 (the string data is stored in the same allocation as the Value of a pair).</p>
1735 <p>StringMap also provides query methods that take byte ranges, so it only ever
1736 copies a string if a value is inserted into the table.</p>
1739 <!-- _______________________________________________________________________ -->
1741 <a name="dss_indexedmap">"llvm/ADT/IndexedMap.h"</a>
1746 IndexedMap is a specialized container for mapping small dense integers (or
1747 values that can be mapped to small dense integers) to some other type. It is
1748 internally implemented as a vector with a mapping function that maps the keys to
1749 the dense integer range.
1753 This is useful for cases like virtual registers in the LLVM code generator: they
1754 have a dense mapping that is offset by a compile-time constant (the first
1755 virtual register ID).</p>
1759 <!-- _______________________________________________________________________ -->
1761 <a name="dss_densemap">"llvm/ADT/DenseMap.h"</a>
1767 DenseMap is a simple quadratically probed hash table. It excels at supporting
1768 small keys and values: it uses a single allocation to hold all of the pairs that
1769 are currently inserted in the map. DenseMap is a great way to map pointers to
1770 pointers, or map other small types to each other.
1774 There are several aspects of DenseMap that you should be aware of, however. The
1775 iterators in a DenseMap are invalidated whenever an insertion occurs, unlike
1776 map. Also, because DenseMap allocates space for a large number of key/value
1777 pairs (it starts with 64 by default), it will waste a lot of space if your keys
1778 or values are large. Finally, you must implement a partial specialization of
1779 DenseMapInfo for the key that you want, if it isn't already supported. This
1780 is required to tell DenseMap about two special marker values (which can never be
1781 inserted into the map) that it needs internally.</p>
1784 DenseMap's find_as() method supports lookup operations using an alternate key
1785 type. This is useful in cases where the normal key type is expensive to
1786 construct, but cheap to compare against. The DenseMapInfo is responsible for
1787 defining the appropriate comparison and hashing methods for each alternate
1793 <!-- _______________________________________________________________________ -->
1795 <a name="dss_valuemap">"llvm/ADT/ValueMap.h"</a>
1801 ValueMap is a wrapper around a <a href="#dss_densemap">DenseMap</a> mapping
1802 Value*s (or subclasses) to another type. When a Value is deleted or RAUW'ed,
1803 ValueMap will update itself so the new version of the key is mapped to the same
1804 value, just as if the key were a WeakVH. You can configure exactly how this
1805 happens, and what else happens on these two events, by passing
1806 a <code>Config</code> parameter to the ValueMap template.</p>
1810 <!-- _______________________________________________________________________ -->
1812 <a name="dss_intervalmap">"llvm/ADT/IntervalMap.h"</a>
1817 <p> IntervalMap is a compact map for small keys and values. It maps key
1818 intervals instead of single keys, and it will automatically coalesce adjacent
1819 intervals. When then map only contains a few intervals, they are stored in the
1820 map object itself to avoid allocations.</p>
1822 <p> The IntervalMap iterators are quite big, so they should not be passed around
1823 as STL iterators. The heavyweight iterators allow a smaller data structure.</p>
1827 <!-- _______________________________________________________________________ -->
1829 <a name="dss_map"><map></a>
1835 std::map has similar characteristics to <a href="#dss_set">std::set</a>: it uses
1836 a single allocation per pair inserted into the map, it offers log(n) lookup with
1837 an extremely large constant factor, imposes a space penalty of 3 pointers per
1838 pair in the map, etc.</p>
1840 <p>std::map is most useful when your keys or values are very large, if you need
1841 to iterate over the collection in sorted order, or if you need stable iterators
1842 into the map (i.e. they don't get invalidated if an insertion or deletion of
1843 another element takes place).</p>
1847 <!-- _______________________________________________________________________ -->
1849 <a name="dss_inteqclasses">"llvm/ADT/IntEqClasses.h"</a>
1854 <p>IntEqClasses provides a compact representation of equivalence classes of
1855 small integers. Initially, each integer in the range 0..n-1 has its own
1856 equivalence class. Classes can be joined by passing two class representatives to
1857 the join(a, b) method. Two integers are in the same class when findLeader()
1858 returns the same representative.</p>
1860 <p>Once all equivalence classes are formed, the map can be compressed so each
1861 integer 0..n-1 maps to an equivalence class number in the range 0..m-1, where m
1862 is the total number of equivalence classes. The map must be uncompressed before
1863 it can be edited again.</p>
1867 <!-- _______________________________________________________________________ -->
1869 <a name="dss_immutablemap">"llvm/ADT/ImmutableMap.h"</a>
1875 ImmutableMap is an immutable (functional) map implementation based on an AVL
1877 Adding or removing elements is done through a Factory object and results in the
1878 creation of a new ImmutableMap object.
1879 If an ImmutableMap already exists with the given key set, then the existing one
1880 is returned; equality is compared with a FoldingSetNodeID.
1881 The time and space complexity of add or remove operations is logarithmic in the
1882 size of the original map.
1886 <!-- _______________________________________________________________________ -->
1888 <a name="dss_othermap">Other Map-Like Container Options</a>
1894 The STL provides several other options, such as std::multimap and the various
1895 "hash_map" like containers (whether from C++ TR1 or from the SGI library). We
1896 never use hash_set and unordered_set because they are generally very expensive
1897 (each insertion requires a malloc) and very non-portable.</p>
1899 <p>std::multimap is useful if you want to map a key to multiple values, but has
1900 all the drawbacks of std::map. A sorted vector or some other approach is almost
1907 <!-- ======================================================================= -->
1909 <a name="ds_bit">Bit storage containers (BitVector, SparseBitVector)</a>
1913 <p>Unlike the other containers, there are only two bit storage containers, and
1914 choosing when to use each is relatively straightforward.</p>
1916 <p>One additional option is
1917 <tt>std::vector<bool></tt>: we discourage its use for two reasons 1) the
1918 implementation in many common compilers (e.g. commonly available versions of
1919 GCC) is extremely inefficient and 2) the C++ standards committee is likely to
1920 deprecate this container and/or change it significantly somehow. In any case,
1921 please don't use it.</p>
1923 <!-- _______________________________________________________________________ -->
1925 <a name="dss_bitvector">BitVector</a>
1929 <p> The BitVector container provides a dynamic size set of bits for manipulation.
1930 It supports individual bit setting/testing, as well as set operations. The set
1931 operations take time O(size of bitvector), but operations are performed one word
1932 at a time, instead of one bit at a time. This makes the BitVector very fast for
1933 set operations compared to other containers. Use the BitVector when you expect
1934 the number of set bits to be high (IE a dense set).
1938 <!-- _______________________________________________________________________ -->
1940 <a name="dss_smallbitvector">SmallBitVector</a>
1944 <p> The SmallBitVector container provides the same interface as BitVector, but
1945 it is optimized for the case where only a small number of bits, less than
1946 25 or so, are needed. It also transparently supports larger bit counts, but
1947 slightly less efficiently than a plain BitVector, so SmallBitVector should
1948 only be used when larger counts are rare.
1952 At this time, SmallBitVector does not support set operations (and, or, xor),
1953 and its operator[] does not provide an assignable lvalue.
1957 <!-- _______________________________________________________________________ -->
1959 <a name="dss_sparsebitvector">SparseBitVector</a>
1963 <p> The SparseBitVector container is much like BitVector, with one major
1964 difference: Only the bits that are set, are stored. This makes the
1965 SparseBitVector much more space efficient than BitVector when the set is sparse,
1966 as well as making set operations O(number of set bits) instead of O(size of
1967 universe). The downside to the SparseBitVector is that setting and testing of random bits is O(N), and on large SparseBitVectors, this can be slower than BitVector. In our implementation, setting or testing bits in sorted order
1968 (either forwards or reverse) is O(1) worst case. Testing and setting bits within 128 bits (depends on size) of the current bit is also O(1). As a general statement, testing/setting bits in a SparseBitVector is O(distance away from last set bit).
1976 <!-- *********************************************************************** -->
1978 <a name="common">Helpful Hints for Common Operations</a>
1980 <!-- *********************************************************************** -->
1984 <p>This section describes how to perform some very simple transformations of
1985 LLVM code. This is meant to give examples of common idioms used, showing the
1986 practical side of LLVM transformations. <p> Because this is a "how-to" section,
1987 you should also read about the main classes that you will be working with. The
1988 <a href="#coreclasses">Core LLVM Class Hierarchy Reference</a> contains details
1989 and descriptions of the main classes that you should know about.</p>
1991 <!-- NOTE: this section should be heavy on example code -->
1992 <!-- ======================================================================= -->
1994 <a name="inspection">Basic Inspection and Traversal Routines</a>
1999 <p>The LLVM compiler infrastructure have many different data structures that may
2000 be traversed. Following the example of the C++ standard template library, the
2001 techniques used to traverse these various data structures are all basically the
2002 same. For a enumerable sequence of values, the <tt>XXXbegin()</tt> function (or
2003 method) returns an iterator to the start of the sequence, the <tt>XXXend()</tt>
2004 function returns an iterator pointing to one past the last valid element of the
2005 sequence, and there is some <tt>XXXiterator</tt> data type that is common
2006 between the two operations.</p>
2008 <p>Because the pattern for iteration is common across many different aspects of
2009 the program representation, the standard template library algorithms may be used
2010 on them, and it is easier to remember how to iterate. First we show a few common
2011 examples of the data structures that need to be traversed. Other data
2012 structures are traversed in very similar ways.</p>
2014 <!-- _______________________________________________________________________ -->
2016 <a name="iterate_function">Iterating over the </a><a
2017 href="#BasicBlock"><tt>BasicBlock</tt></a>s in a <a
2018 href="#Function"><tt>Function</tt></a>
2023 <p>It's quite common to have a <tt>Function</tt> instance that you'd like to
2024 transform in some way; in particular, you'd like to manipulate its
2025 <tt>BasicBlock</tt>s. To facilitate this, you'll need to iterate over all of
2026 the <tt>BasicBlock</tt>s that constitute the <tt>Function</tt>. The following is
2027 an example that prints the name of a <tt>BasicBlock</tt> and the number of
2028 <tt>Instruction</tt>s it contains:</p>
2030 <div class="doc_code">
2032 // <i>func is a pointer to a Function instance</i>
2033 for (Function::iterator i = func->begin(), e = func->end(); i != e; ++i)
2034 // <i>Print out the name of the basic block if it has one, and then the</i>
2035 // <i>number of instructions that it contains</i>
2036 errs() << "Basic block (name=" << i->getName() << ") has "
2037 << i->size() << " instructions.\n";
2041 <p>Note that i can be used as if it were a pointer for the purposes of
2042 invoking member functions of the <tt>Instruction</tt> class. This is
2043 because the indirection operator is overloaded for the iterator
2044 classes. In the above code, the expression <tt>i->size()</tt> is
2045 exactly equivalent to <tt>(*i).size()</tt> just like you'd expect.</p>
2049 <!-- _______________________________________________________________________ -->
2051 <a name="iterate_basicblock">Iterating over the </a><a
2052 href="#Instruction"><tt>Instruction</tt></a>s in a <a
2053 href="#BasicBlock"><tt>BasicBlock</tt></a>
2058 <p>Just like when dealing with <tt>BasicBlock</tt>s in <tt>Function</tt>s, it's
2059 easy to iterate over the individual instructions that make up
2060 <tt>BasicBlock</tt>s. Here's a code snippet that prints out each instruction in
2061 a <tt>BasicBlock</tt>:</p>
2063 <div class="doc_code">
2065 // <i>blk is a pointer to a BasicBlock instance</i>
2066 for (BasicBlock::iterator i = blk->begin(), e = blk->end(); i != e; ++i)
2067 // <i>The next statement works since operator<<(ostream&,...)</i>
2068 // <i>is overloaded for Instruction&</i>
2069 errs() << *i << "\n";
2073 <p>However, this isn't really the best way to print out the contents of a
2074 <tt>BasicBlock</tt>! Since the ostream operators are overloaded for virtually
2075 anything you'll care about, you could have just invoked the print routine on the
2076 basic block itself: <tt>errs() << *blk << "\n";</tt>.</p>
2080 <!-- _______________________________________________________________________ -->
2082 <a name="iterate_institer">Iterating over the </a><a
2083 href="#Instruction"><tt>Instruction</tt></a>s in a <a
2084 href="#Function"><tt>Function</tt></a>
2089 <p>If you're finding that you commonly iterate over a <tt>Function</tt>'s
2090 <tt>BasicBlock</tt>s and then that <tt>BasicBlock</tt>'s <tt>Instruction</tt>s,
2091 <tt>InstIterator</tt> should be used instead. You'll need to include <a
2092 href="/doxygen/InstIterator_8h-source.html"><tt>llvm/Support/InstIterator.h</tt></a>,
2093 and then instantiate <tt>InstIterator</tt>s explicitly in your code. Here's a
2094 small example that shows how to dump all instructions in a function to the standard error stream:<p>
2096 <div class="doc_code">
2098 #include "<a href="/doxygen/InstIterator_8h-source.html">llvm/Support/InstIterator.h</a>"
2100 // <i>F is a pointer to a Function instance</i>
2101 for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
2102 errs() << *I << "\n";
2106 <p>Easy, isn't it? You can also use <tt>InstIterator</tt>s to fill a
2107 work list with its initial contents. For example, if you wanted to
2108 initialize a work list to contain all instructions in a <tt>Function</tt>
2109 F, all you would need to do is something like:</p>
2111 <div class="doc_code">
2113 std::set<Instruction*> worklist;
2114 // or better yet, SmallPtrSet<Instruction*, 64> worklist;
2116 for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
2117 worklist.insert(&*I);
2121 <p>The STL set <tt>worklist</tt> would now contain all instructions in the
2122 <tt>Function</tt> pointed to by F.</p>
2126 <!-- _______________________________________________________________________ -->
2128 <a name="iterate_convert">Turning an iterator into a class pointer (and
2134 <p>Sometimes, it'll be useful to grab a reference (or pointer) to a class
2135 instance when all you've got at hand is an iterator. Well, extracting
2136 a reference or a pointer from an iterator is very straight-forward.
2137 Assuming that <tt>i</tt> is a <tt>BasicBlock::iterator</tt> and <tt>j</tt>
2138 is a <tt>BasicBlock::const_iterator</tt>:</p>
2140 <div class="doc_code">
2142 Instruction& inst = *i; // <i>Grab reference to instruction reference</i>
2143 Instruction* pinst = &*i; // <i>Grab pointer to instruction reference</i>
2144 const Instruction& inst = *j;
2148 <p>However, the iterators you'll be working with in the LLVM framework are
2149 special: they will automatically convert to a ptr-to-instance type whenever they
2150 need to. Instead of dereferencing the iterator and then taking the address of
2151 the result, you can simply assign the iterator to the proper pointer type and
2152 you get the dereference and address-of operation as a result of the assignment
2153 (behind the scenes, this is a result of overloading casting mechanisms). Thus
2154 the last line of the last example,</p>
2156 <div class="doc_code">
2158 Instruction *pinst = &*i;
2162 <p>is semantically equivalent to</p>
2164 <div class="doc_code">
2166 Instruction *pinst = i;
2170 <p>It's also possible to turn a class pointer into the corresponding iterator,
2171 and this is a constant time operation (very efficient). The following code
2172 snippet illustrates use of the conversion constructors provided by LLVM
2173 iterators. By using these, you can explicitly grab the iterator of something
2174 without actually obtaining it via iteration over some structure:</p>
2176 <div class="doc_code">
2178 void printNextInstruction(Instruction* inst) {
2179 BasicBlock::iterator it(inst);
2180 ++it; // <i>After this line, it refers to the instruction after *inst</i>
2181 if (it != inst->getParent()->end()) errs() << *it << "\n";
2186 <p>Unfortunately, these implicit conversions come at a cost; they prevent
2187 these iterators from conforming to standard iterator conventions, and thus
2188 from being usable with standard algorithms and containers. For example, they
2189 prevent the following code, where <tt>B</tt> is a <tt>BasicBlock</tt>,
2192 <div class="doc_code">
2194 llvm::SmallVector<llvm::Instruction *, 16>(B->begin(), B->end());
2198 <p>Because of this, these implicit conversions may be removed some day,
2199 and <tt>operator*</tt> changed to return a pointer instead of a reference.</p>
2203 <!--_______________________________________________________________________-->
2205 <a name="iterate_complex">Finding call sites: a slightly more complex
2211 <p>Say that you're writing a FunctionPass and would like to count all the
2212 locations in the entire module (that is, across every <tt>Function</tt>) where a
2213 certain function (i.e., some <tt>Function</tt>*) is already in scope. As you'll
2214 learn later, you may want to use an <tt>InstVisitor</tt> to accomplish this in a
2215 much more straight-forward manner, but this example will allow us to explore how
2216 you'd do it if you didn't have <tt>InstVisitor</tt> around. In pseudo-code, this
2217 is what we want to do:</p>
2219 <div class="doc_code">
2221 initialize callCounter to zero
2222 for each Function f in the Module
2223 for each BasicBlock b in f
2224 for each Instruction i in b
2225 if (i is a CallInst and calls the given function)
2226 increment callCounter
2230 <p>And the actual code is (remember, because we're writing a
2231 <tt>FunctionPass</tt>, our <tt>FunctionPass</tt>-derived class simply has to
2232 override the <tt>runOnFunction</tt> method):</p>
2234 <div class="doc_code">
2236 Function* targetFunc = ...;
2238 class OurFunctionPass : public FunctionPass {
2240 OurFunctionPass(): callCounter(0) { }
2242 virtual runOnFunction(Function& F) {
2243 for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
2244 for (BasicBlock::iterator i = b->begin(), ie = b->end(); i != ie; ++i) {
2245 if (<a href="#CallInst">CallInst</a>* callInst = <a href="#isa">dyn_cast</a><<a
2246 href="#CallInst">CallInst</a>>(&*i)) {
2247 // <i>We know we've encountered a call instruction, so we</i>
2248 // <i>need to determine if it's a call to the</i>
2249 // <i>function pointed to by m_func or not.</i>
2250 if (callInst->getCalledFunction() == targetFunc)
2258 unsigned callCounter;
2265 <!--_______________________________________________________________________-->
2267 <a name="calls_and_invokes">Treating calls and invokes the same way</a>
2272 <p>You may have noticed that the previous example was a bit oversimplified in
2273 that it did not deal with call sites generated by 'invoke' instructions. In
2274 this, and in other situations, you may find that you want to treat
2275 <tt>CallInst</tt>s and <tt>InvokeInst</tt>s the same way, even though their
2276 most-specific common base class is <tt>Instruction</tt>, which includes lots of
2277 less closely-related things. For these cases, LLVM provides a handy wrapper
2279 href="http://llvm.org/doxygen/classllvm_1_1CallSite.html"><tt>CallSite</tt></a>.
2280 It is essentially a wrapper around an <tt>Instruction</tt> pointer, with some
2281 methods that provide functionality common to <tt>CallInst</tt>s and
2282 <tt>InvokeInst</tt>s.</p>
2284 <p>This class has "value semantics": it should be passed by value, not by
2285 reference and it should not be dynamically allocated or deallocated using
2286 <tt>operator new</tt> or <tt>operator delete</tt>. It is efficiently copyable,
2287 assignable and constructable, with costs equivalents to that of a bare pointer.
2288 If you look at its definition, it has only a single pointer member.</p>
2292 <!--_______________________________________________________________________-->
2294 <a name="iterate_chains">Iterating over def-use & use-def chains</a>
2299 <p>Frequently, we might have an instance of the <a
2300 href="/doxygen/classllvm_1_1Value.html">Value Class</a> and we want to
2301 determine which <tt>User</tt>s use the <tt>Value</tt>. The list of all
2302 <tt>User</tt>s of a particular <tt>Value</tt> is called a <i>def-use</i> chain.
2303 For example, let's say we have a <tt>Function*</tt> named <tt>F</tt> to a
2304 particular function <tt>foo</tt>. Finding all of the instructions that
2305 <i>use</i> <tt>foo</tt> is as simple as iterating over the <i>def-use</i> chain
2308 <div class="doc_code">
2312 for (Value::use_iterator i = F->use_begin(), e = F->use_end(); i != e; ++i)
2313 if (Instruction *Inst = dyn_cast<Instruction>(*i)) {
2314 errs() << "F is used in instruction:\n";
2315 errs() << *Inst << "\n";
2320 <p>Note that dereferencing a <tt>Value::use_iterator</tt> is not a very cheap
2321 operation. Instead of performing <tt>*i</tt> above several times, consider
2322 doing it only once in the loop body and reusing its result.</p>
2324 <p>Alternatively, it's common to have an instance of the <a
2325 href="/doxygen/classllvm_1_1User.html">User Class</a> and need to know what
2326 <tt>Value</tt>s are used by it. The list of all <tt>Value</tt>s used by a
2327 <tt>User</tt> is known as a <i>use-def</i> chain. Instances of class
2328 <tt>Instruction</tt> are common <tt>User</tt>s, so we might want to iterate over
2329 all of the values that a particular instruction uses (that is, the operands of
2330 the particular <tt>Instruction</tt>):</p>
2332 <div class="doc_code">
2334 Instruction *pi = ...;
2336 for (User::op_iterator i = pi->op_begin(), e = pi->op_end(); i != e; ++i) {
2343 <p>Declaring objects as <tt>const</tt> is an important tool of enforcing
2344 mutation free algorithms (such as analyses, etc.). For this purpose above
2345 iterators come in constant flavors as <tt>Value::const_use_iterator</tt>
2346 and <tt>Value::const_op_iterator</tt>. They automatically arise when
2347 calling <tt>use/op_begin()</tt> on <tt>const Value*</tt>s or
2348 <tt>const User*</tt>s respectively. Upon dereferencing, they return
2349 <tt>const Use*</tt>s. Otherwise the above patterns remain unchanged.</p>
2353 <!--_______________________________________________________________________-->
2355 <a name="iterate_preds">Iterating over predecessors &
2356 successors of blocks</a>
2361 <p>Iterating over the predecessors and successors of a block is quite easy
2362 with the routines defined in <tt>"llvm/Support/CFG.h"</tt>. Just use code like
2363 this to iterate over all predecessors of BB:</p>
2365 <div class="doc_code">
2367 #include "llvm/Support/CFG.h"
2368 BasicBlock *BB = ...;
2370 for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
2371 BasicBlock *Pred = *PI;
2377 <p>Similarly, to iterate over successors use
2378 succ_iterator/succ_begin/succ_end.</p>
2384 <!-- ======================================================================= -->
2386 <a name="simplechanges">Making simple changes</a>
2391 <p>There are some primitive transformation operations present in the LLVM
2392 infrastructure that are worth knowing about. When performing
2393 transformations, it's fairly common to manipulate the contents of basic
2394 blocks. This section describes some of the common methods for doing so
2395 and gives example code.</p>
2397 <!--_______________________________________________________________________-->
2399 <a name="schanges_creating">Creating and inserting new
2400 <tt>Instruction</tt>s</a>
2405 <p><i>Instantiating Instructions</i></p>
2407 <p>Creation of <tt>Instruction</tt>s is straight-forward: simply call the
2408 constructor for the kind of instruction to instantiate and provide the necessary
2409 parameters. For example, an <tt>AllocaInst</tt> only <i>requires</i> a
2410 (const-ptr-to) <tt>Type</tt>. Thus:</p>
2412 <div class="doc_code">
2414 AllocaInst* ai = new AllocaInst(Type::Int32Ty);
2418 <p>will create an <tt>AllocaInst</tt> instance that represents the allocation of
2419 one integer in the current stack frame, at run time. Each <tt>Instruction</tt>
2420 subclass is likely to have varying default parameters which change the semantics
2421 of the instruction, so refer to the <a
2422 href="/doxygen/classllvm_1_1Instruction.html">doxygen documentation for the subclass of
2423 Instruction</a> that you're interested in instantiating.</p>
2425 <p><i>Naming values</i></p>
2427 <p>It is very useful to name the values of instructions when you're able to, as
2428 this facilitates the debugging of your transformations. If you end up looking
2429 at generated LLVM machine code, you definitely want to have logical names
2430 associated with the results of instructions! By supplying a value for the
2431 <tt>Name</tt> (default) parameter of the <tt>Instruction</tt> constructor, you
2432 associate a logical name with the result of the instruction's execution at
2433 run time. For example, say that I'm writing a transformation that dynamically
2434 allocates space for an integer on the stack, and that integer is going to be
2435 used as some kind of index by some other code. To accomplish this, I place an
2436 <tt>AllocaInst</tt> at the first point in the first <tt>BasicBlock</tt> of some
2437 <tt>Function</tt>, and I'm intending to use it within the same
2438 <tt>Function</tt>. I might do:</p>
2440 <div class="doc_code">
2442 AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
2446 <p>where <tt>indexLoc</tt> is now the logical name of the instruction's
2447 execution value, which is a pointer to an integer on the run time stack.</p>
2449 <p><i>Inserting instructions</i></p>
2451 <p>There are essentially two ways to insert an <tt>Instruction</tt>
2452 into an existing sequence of instructions that form a <tt>BasicBlock</tt>:</p>
2455 <li>Insertion into an explicit instruction list
2457 <p>Given a <tt>BasicBlock* pb</tt>, an <tt>Instruction* pi</tt> within that
2458 <tt>BasicBlock</tt>, and a newly-created instruction we wish to insert
2459 before <tt>*pi</tt>, we do the following: </p>
2461 <div class="doc_code">
2463 BasicBlock *pb = ...;
2464 Instruction *pi = ...;
2465 Instruction *newInst = new Instruction(...);
2467 pb->getInstList().insert(pi, newInst); // <i>Inserts newInst before pi in pb</i>
2471 <p>Appending to the end of a <tt>BasicBlock</tt> is so common that
2472 the <tt>Instruction</tt> class and <tt>Instruction</tt>-derived
2473 classes provide constructors which take a pointer to a
2474 <tt>BasicBlock</tt> to be appended to. For example code that
2477 <div class="doc_code">
2479 BasicBlock *pb = ...;
2480 Instruction *newInst = new Instruction(...);
2482 pb->getInstList().push_back(newInst); // <i>Appends newInst to pb</i>
2488 <div class="doc_code">
2490 BasicBlock *pb = ...;
2491 Instruction *newInst = new Instruction(..., pb);
2495 <p>which is much cleaner, especially if you are creating
2496 long instruction streams.</p></li>
2498 <li>Insertion into an implicit instruction list
2500 <p><tt>Instruction</tt> instances that are already in <tt>BasicBlock</tt>s
2501 are implicitly associated with an existing instruction list: the instruction
2502 list of the enclosing basic block. Thus, we could have accomplished the same
2503 thing as the above code without being given a <tt>BasicBlock</tt> by doing:
2506 <div class="doc_code">
2508 Instruction *pi = ...;
2509 Instruction *newInst = new Instruction(...);
2511 pi->getParent()->getInstList().insert(pi, newInst);
2515 <p>In fact, this sequence of steps occurs so frequently that the
2516 <tt>Instruction</tt> class and <tt>Instruction</tt>-derived classes provide
2517 constructors which take (as a default parameter) a pointer to an
2518 <tt>Instruction</tt> which the newly-created <tt>Instruction</tt> should
2519 precede. That is, <tt>Instruction</tt> constructors are capable of
2520 inserting the newly-created instance into the <tt>BasicBlock</tt> of a
2521 provided instruction, immediately before that instruction. Using an
2522 <tt>Instruction</tt> constructor with a <tt>insertBefore</tt> (default)
2523 parameter, the above code becomes:</p>
2525 <div class="doc_code">
2527 Instruction* pi = ...;
2528 Instruction* newInst = new Instruction(..., pi);
2532 <p>which is much cleaner, especially if you're creating a lot of
2533 instructions and adding them to <tt>BasicBlock</tt>s.</p></li>
2538 <!--_______________________________________________________________________-->
2540 <a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a>
2545 <p>Deleting an instruction from an existing sequence of instructions that form a
2546 <a href="#BasicBlock"><tt>BasicBlock</tt></a> is very straight-forward: just
2547 call the instruction's eraseFromParent() method. For example:</p>
2549 <div class="doc_code">
2551 <a href="#Instruction">Instruction</a> *I = .. ;
2552 I->eraseFromParent();
2556 <p>This unlinks the instruction from its containing basic block and deletes
2557 it. If you'd just like to unlink the instruction from its containing basic
2558 block but not delete it, you can use the <tt>removeFromParent()</tt> method.</p>
2562 <!--_______________________________________________________________________-->
2564 <a name="schanges_replacing">Replacing an <tt>Instruction</tt> with another
2570 <p><i>Replacing individual instructions</i></p>
2572 <p>Including "<a href="/doxygen/BasicBlockUtils_8h-source.html">llvm/Transforms/Utils/BasicBlockUtils.h</a>"
2573 permits use of two very useful replace functions: <tt>ReplaceInstWithValue</tt>
2574 and <tt>ReplaceInstWithInst</tt>.</p>
2576 <h5><a name="schanges_deleting">Deleting <tt>Instruction</tt>s</a></h5>
2579 <li><tt>ReplaceInstWithValue</tt>
2581 <p>This function replaces all uses of a given instruction with a value,
2582 and then removes the original instruction. The following example
2583 illustrates the replacement of the result of a particular
2584 <tt>AllocaInst</tt> that allocates memory for a single integer with a null
2585 pointer to an integer.</p>
2587 <div class="doc_code">
2589 AllocaInst* instToReplace = ...;
2590 BasicBlock::iterator ii(instToReplace);
2592 ReplaceInstWithValue(instToReplace->getParent()->getInstList(), ii,
2593 Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
2596 <li><tt>ReplaceInstWithInst</tt>
2598 <p>This function replaces a particular instruction with another
2599 instruction, inserting the new instruction into the basic block at the
2600 location where the old instruction was, and replacing any uses of the old
2601 instruction with the new instruction. The following example illustrates
2602 the replacement of one <tt>AllocaInst</tt> with another.</p>
2604 <div class="doc_code">
2606 AllocaInst* instToReplace = ...;
2607 BasicBlock::iterator ii(instToReplace);
2609 ReplaceInstWithInst(instToReplace->getParent()->getInstList(), ii,
2610 new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
2614 <p><i>Replacing multiple uses of <tt>User</tt>s and <tt>Value</tt>s</i></p>
2616 <p>You can use <tt>Value::replaceAllUsesWith</tt> and
2617 <tt>User::replaceUsesOfWith</tt> to change more than one use at a time. See the
2618 doxygen documentation for the <a href="/doxygen/classllvm_1_1Value.html">Value Class</a>
2619 and <a href="/doxygen/classllvm_1_1User.html">User Class</a>, respectively, for more
2622 <!-- Value::replaceAllUsesWith User::replaceUsesOfWith Point out:
2623 include/llvm/Transforms/Utils/ especially BasicBlockUtils.h with:
2624 ReplaceInstWithValue, ReplaceInstWithInst -->
2628 <!--_______________________________________________________________________-->
2630 <a name="schanges_deletingGV">Deleting <tt>GlobalVariable</tt>s</a>
2635 <p>Deleting a global variable from a module is just as easy as deleting an
2636 Instruction. First, you must have a pointer to the global variable that you wish
2637 to delete. You use this pointer to erase it from its parent, the module.
2640 <div class="doc_code">
2642 <a href="#GlobalVariable">GlobalVariable</a> *GV = .. ;
2644 GV->eraseFromParent();
2652 <!-- ======================================================================= -->
2654 <a name="create_types">How to Create Types</a>
2659 <p>In generating IR, you may need some complex types. If you know these types
2660 statically, you can use <tt>TypeBuilder<...>::get()</tt>, defined
2661 in <tt>llvm/Support/TypeBuilder.h</tt>, to retrieve them. <tt>TypeBuilder</tt>
2662 has two forms depending on whether you're building types for cross-compilation
2663 or native library use. <tt>TypeBuilder<T, true></tt> requires
2664 that <tt>T</tt> be independent of the host environment, meaning that it's built
2666 the <a href="/doxygen/namespacellvm_1_1types.html"><tt>llvm::types</tt></a>
2667 namespace and pointers, functions, arrays, etc. built of
2668 those. <tt>TypeBuilder<T, false></tt> additionally allows native C types
2669 whose size may depend on the host compiler. For example,</p>
2671 <div class="doc_code">
2673 FunctionType *ft = TypeBuilder<types::i<8>(types::i<32>*), true>::get();
2677 <p>is easier to read and write than the equivalent</p>
2679 <div class="doc_code">
2681 std::vector<const Type*> params;
2682 params.push_back(PointerType::getUnqual(Type::Int32Ty));
2683 FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
2687 <p>See the <a href="/doxygen/TypeBuilder_8h-source.html#l00001">class
2688 comment</a> for more details.</p>
2694 <!-- *********************************************************************** -->
2696 <a name="threading">Threads and LLVM</a>
2698 <!-- *********************************************************************** -->
2702 This section describes the interaction of the LLVM APIs with multithreading,
2703 both on the part of client applications, and in the JIT, in the hosted
2708 Note that LLVM's support for multithreading is still relatively young. Up
2709 through version 2.5, the execution of threaded hosted applications was
2710 supported, but not threaded client access to the APIs. While this use case is
2711 now supported, clients <em>must</em> adhere to the guidelines specified below to
2712 ensure proper operation in multithreaded mode.
2716 Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
2717 intrinsics in order to support threaded operation. If you need a
2718 multhreading-capable LLVM on a platform without a suitably modern system
2719 compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
2720 using the resultant compiler to build a copy of LLVM with multithreading
2724 <!-- ======================================================================= -->
2726 <a name="startmultithreaded">Entering and Exiting Multithreaded Mode</a>
2732 In order to properly protect its internal data structures while avoiding
2733 excessive locking overhead in the single-threaded case, the LLVM must intialize
2734 certain data structures necessary to provide guards around its internals. To do
2735 so, the client program must invoke <tt>llvm_start_multithreaded()</tt> before
2736 making any concurrent LLVM API calls. To subsequently tear down these
2737 structures, use the <tt>llvm_stop_multithreaded()</tt> call. You can also use
2738 the <tt>llvm_is_multithreaded()</tt> call to check the status of multithreaded
2743 Note that both of these calls must be made <em>in isolation</em>. That is to
2744 say that no other LLVM API calls may be executing at any time during the
2745 execution of <tt>llvm_start_multithreaded()</tt> or <tt>llvm_stop_multithreaded
2746 </tt>. It's is the client's responsibility to enforce this isolation.
2750 The return value of <tt>llvm_start_multithreaded()</tt> indicates the success or
2751 failure of the initialization. Failure typically indicates that your copy of
2752 LLVM was built without multithreading support, typically because GCC atomic
2753 intrinsics were not found in your system compiler. In this case, the LLVM API
2754 will not be safe for concurrent calls. However, it <em>will</em> be safe for
2755 hosting threaded applications in the JIT, though <a href="#jitthreading">care
2756 must be taken</a> to ensure that side exits and the like do not accidentally
2757 result in concurrent LLVM API calls.
2761 <!-- ======================================================================= -->
2763 <a name="shutdown">Ending Execution with <tt>llvm_shutdown()</tt></a>
2768 When you are done using the LLVM APIs, you should call <tt>llvm_shutdown()</tt>
2769 to deallocate memory used for internal structures. This will also invoke
2770 <tt>llvm_stop_multithreaded()</tt> if LLVM is operating in multithreaded mode.
2771 As such, <tt>llvm_shutdown()</tt> requires the same isolation guarantees as
2772 <tt>llvm_stop_multithreaded()</tt>.
2776 Note that, if you use scope-based shutdown, you can use the
2777 <tt>llvm_shutdown_obj</tt> class, which calls <tt>llvm_shutdown()</tt> in its
2781 <!-- ======================================================================= -->
2783 <a name="managedstatic">Lazy Initialization with <tt>ManagedStatic</tt></a>
2788 <tt>ManagedStatic</tt> is a utility class in LLVM used to implement static
2789 initialization of static resources, such as the global type tables. Before the
2790 invocation of <tt>llvm_shutdown()</tt>, it implements a simple lazy
2791 initialization scheme. Once <tt>llvm_start_multithreaded()</tt> returns,
2792 however, it uses double-checked locking to implement thread-safe lazy
2797 Note that, because no other threads are allowed to issue LLVM API calls before
2798 <tt>llvm_start_multithreaded()</tt> returns, it is possible to have
2799 <tt>ManagedStatic</tt>s of <tt>llvm::sys::Mutex</tt>s.
2803 The <tt>llvm_acquire_global_lock()</tt> and <tt>llvm_release_global_lock</tt>
2804 APIs provide access to the global lock used to implement the double-checked
2805 locking for lazy initialization. These should only be used internally to LLVM,
2806 and only if you know what you're doing!
2810 <!-- ======================================================================= -->
2812 <a name="llvmcontext">Achieving Isolation with <tt>LLVMContext</tt></a>
2817 <tt>LLVMContext</tt> is an opaque class in the LLVM API which clients can use
2818 to operate multiple, isolated instances of LLVM concurrently within the same
2819 address space. For instance, in a hypothetical compile-server, the compilation
2820 of an individual translation unit is conceptually independent from all the
2821 others, and it would be desirable to be able to compile incoming translation
2822 units concurrently on independent server threads. Fortunately,
2823 <tt>LLVMContext</tt> exists to enable just this kind of scenario!
2827 Conceptually, <tt>LLVMContext</tt> provides isolation. Every LLVM entity
2828 (<tt>Module</tt>s, <tt>Value</tt>s, <tt>Type</tt>s, <tt>Constant</tt>s, etc.)
2829 in LLVM's in-memory IR belongs to an <tt>LLVMContext</tt>. Entities in
2830 different contexts <em>cannot</em> interact with each other: <tt>Module</tt>s in
2831 different contexts cannot be linked together, <tt>Function</tt>s cannot be added
2832 to <tt>Module</tt>s in different contexts, etc. What this means is that is is
2833 safe to compile on multiple threads simultaneously, as long as no two threads
2834 operate on entities within the same context.
2838 In practice, very few places in the API require the explicit specification of a
2839 <tt>LLVMContext</tt>, other than the <tt>Type</tt> creation/lookup APIs.
2840 Because every <tt>Type</tt> carries a reference to its owning context, most
2841 other entities can determine what context they belong to by looking at their
2842 own <tt>Type</tt>. If you are adding new entities to LLVM IR, please try to
2843 maintain this interface design.
2847 For clients that do <em>not</em> require the benefits of isolation, LLVM
2848 provides a convenience API <tt>getGlobalContext()</tt>. This returns a global,
2849 lazily initialized <tt>LLVMContext</tt> that may be used in situations where
2850 isolation is not a concern.
2854 <!-- ======================================================================= -->
2856 <a name="jitthreading">Threads and the JIT</a>
2861 LLVM's "eager" JIT compiler is safe to use in threaded programs. Multiple
2862 threads can call <tt>ExecutionEngine::getPointerToFunction()</tt> or
2863 <tt>ExecutionEngine::runFunction()</tt> concurrently, and multiple threads can
2864 run code output by the JIT concurrently. The user must still ensure that only
2865 one thread accesses IR in a given <tt>LLVMContext</tt> while another thread
2866 might be modifying it. One way to do that is to always hold the JIT lock while
2867 accessing IR outside the JIT (the JIT <em>modifies</em> the IR by adding
2868 <tt>CallbackVH</tt>s). Another way is to only
2869 call <tt>getPointerToFunction()</tt> from the <tt>LLVMContext</tt>'s thread.
2872 <p>When the JIT is configured to compile lazily (using
2873 <tt>ExecutionEngine::DisableLazyCompilation(false)</tt>), there is currently a
2874 <a href="http://llvm.org/bugs/show_bug.cgi?id=5184">race condition</a> in
2875 updating call sites after a function is lazily-jitted. It's still possible to
2876 use the lazy JIT in a threaded program if you ensure that only one thread at a
2877 time can call any particular lazy stub and that the JIT lock guards any IR
2878 access, but we suggest using only the eager JIT in threaded programs.
2884 <!-- *********************************************************************** -->
2886 <a name="advanced">Advanced Topics</a>
2888 <!-- *********************************************************************** -->
2892 This section describes some of the advanced or obscure API's that most clients
2893 do not need to be aware of. These API's tend manage the inner workings of the
2894 LLVM system, and only need to be accessed in unusual circumstances.
2898 <!-- ======================================================================= -->
2900 <a name="SymbolTable">The <tt>ValueSymbolTable</tt> class</a>
2904 <p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html">
2905 ValueSymbolTable</a></tt> class provides a symbol table that the <a
2906 href="#Function"><tt>Function</tt></a> and <a href="#Module">
2907 <tt>Module</tt></a> classes use for naming value definitions. The symbol table
2908 can provide a name for any <a href="#Value"><tt>Value</tt></a>.
2911 <p>Note that the <tt>SymbolTable</tt> class should not be directly accessed
2912 by most clients. It should only be used when iteration over the symbol table
2913 names themselves are required, which is very special purpose. Note that not
2915 <tt><a href="#Value">Value</a></tt>s have names, and those without names (i.e. they have
2916 an empty name) do not exist in the symbol table.
2919 <p>Symbol tables support iteration over the values in the symbol
2920 table with <tt>begin/end/iterator</tt> and supports querying to see if a
2921 specific name is in the symbol table (with <tt>lookup</tt>). The
2922 <tt>ValueSymbolTable</tt> class exposes no public mutator methods, instead,
2923 simply call <tt>setName</tt> on a value, which will autoinsert it into the
2924 appropriate symbol table.</p>
2930 <!-- ======================================================================= -->
2932 <a name="UserLayout">The <tt>User</tt> and owned <tt>Use</tt> classes' memory layout</a>
2936 <p>The <tt><a href="http://llvm.org/doxygen/classllvm_1_1User.html">
2937 User</a></tt> class provides a basis for expressing the ownership of <tt>User</tt>
2938 towards other <tt><a href="http://llvm.org/doxygen/classllvm_1_1Value.html">
2939 Value</a></tt>s. The <tt><a href="http://llvm.org/doxygen/classllvm_1_1Use.html">
2940 Use</a></tt> helper class is employed to do the bookkeeping and to facilitate <i>O(1)</i>
2941 addition and removal.</p>
2943 <!-- ______________________________________________________________________ -->
2946 Interaction and relationship between <tt>User</tt> and <tt>Use</tt> objects
2952 A subclass of <tt>User</tt> can choose between incorporating its <tt>Use</tt> objects
2953 or refer to them out-of-line by means of a pointer. A mixed variant
2954 (some <tt>Use</tt>s inline others hung off) is impractical and breaks the invariant
2955 that the <tt>Use</tt> objects belonging to the same <tt>User</tt> form a contiguous array.
2959 We have 2 different layouts in the <tt>User</tt> (sub)classes:
2962 The <tt>Use</tt> object(s) are inside (resp. at fixed offset) of the <tt>User</tt>
2963 object and there are a fixed number of them.</p>
2966 The <tt>Use</tt> object(s) are referenced by a pointer to an
2967 array from the <tt>User</tt> object and there may be a variable
2971 As of v2.4 each layout still possesses a direct pointer to the
2972 start of the array of <tt>Use</tt>s. Though not mandatory for layout a),
2973 we stick to this redundancy for the sake of simplicity.
2974 The <tt>User</tt> object also stores the number of <tt>Use</tt> objects it
2975 has. (Theoretically this information can also be calculated
2976 given the scheme presented below.)</p>
2978 Special forms of allocation operators (<tt>operator new</tt>)
2979 enforce the following memory layouts:</p>
2982 <li><p>Layout a) is modelled by prepending the <tt>User</tt> object by the <tt>Use[]</tt> array.</p>
2985 ...---.---.---.---.-------...
2986 | P | P | P | P | User
2987 '''---'---'---'---'-------'''
2990 <li><p>Layout b) is modelled by pointing at the <tt>Use[]</tt> array.</p>
3002 <i>(In the above figures '<tt>P</tt>' stands for the <tt>Use**</tt> that
3003 is stored in each <tt>Use</tt> object in the member <tt>Use::Prev</tt>)</i>
3007 <!-- ______________________________________________________________________ -->
3009 <a name="Waymarking">The waymarking algorithm</a>
3014 Since the <tt>Use</tt> objects are deprived of the direct (back)pointer to
3015 their <tt>User</tt> objects, there must be a fast and exact method to
3016 recover it. This is accomplished by the following scheme:</p>
3018 A bit-encoding in the 2 LSBits (least significant bits) of the <tt>Use::Prev</tt> allows to find the
3019 start of the <tt>User</tt> object:
3021 <li><tt>00</tt> —> binary digit 0</li>
3022 <li><tt>01</tt> —> binary digit 1</li>
3023 <li><tt>10</tt> —> stop and calculate (<tt>s</tt>)</li>
3024 <li><tt>11</tt> —> full stop (<tt>S</tt>)</li>
3027 Given a <tt>Use*</tt>, all we have to do is to walk till we get
3028 a stop and we either have a <tt>User</tt> immediately behind or
3029 we have to walk to the next stop picking up digits
3030 and calculating the offset:</p>
3032 .---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
3033 | 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
3034 '---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
3035 |+15 |+10 |+6 |+3 |+1
3038 | | |______________________>
3039 | |______________________________________>
3040 |__________________________________________________________>
3043 Only the significant number of bits need to be stored between the
3044 stops, so that the <i>worst case is 20 memory accesses</i> when there are
3045 1000 <tt>Use</tt> objects associated with a <tt>User</tt>.</p>
3049 <!-- ______________________________________________________________________ -->
3051 <a name="ReferenceImpl">Reference implementation</a>
3056 The following literate Haskell fragment demonstrates the concept:</p>
3058 <div class="doc_code">
3060 > import Test.QuickCheck
3062 > digits :: Int -> [Char] -> [Char]
3063 > digits 0 acc = '0' : acc
3064 > digits 1 acc = '1' : acc
3065 > digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
3067 > dist :: Int -> [Char] -> [Char]
3070 > dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
3071 > dist n acc = dist (n - 1) $ dist 1 acc
3073 > takeLast n ss = reverse $ take n $ reverse ss
3075 > test = takeLast 40 $ dist 20 []
3080 Printing <test> gives: <tt>"1s100000s11010s10100s1111s1010s110s11s1S"</tt></p>
3082 The reverse algorithm computes the length of the string just by examining
3083 a certain prefix:</p>
3085 <div class="doc_code">
3087 > pref :: [Char] -> Int
3089 > pref ('s':'1':rest) = decode 2 1 rest
3090 > pref (_:rest) = 1 + pref rest
3092 > decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
3093 > decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
3094 > decode walk acc _ = walk + acc
3099 Now, as expected, printing <pref test> gives <tt>40</tt>.</p>
3101 We can <i>quickCheck</i> this with following property:</p>
3103 <div class="doc_code">
3105 > testcase = dist 2000 []
3106 > testcaseLength = length testcase
3108 > identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
3109 > where arr = takeLast n testcase
3114 As expected <quickCheck identityProp> gives:</p>
3117 *Main> quickCheck identityProp
3118 OK, passed 100 tests.
3121 Let's be a bit more exhaustive:</p>
3123 <div class="doc_code">
3126 > deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
3131 And here is the result of <deepCheck identityProp>:</p>
3134 *Main> deepCheck identityProp
3135 OK, passed 500 tests.
3140 <!-- ______________________________________________________________________ -->
3142 <a name="Tagging">Tagging considerations</a>
3148 To maintain the invariant that the 2 LSBits of each <tt>Use**</tt> in <tt>Use</tt>
3149 never change after being set up, setters of <tt>Use::Prev</tt> must re-tag the
3150 new <tt>Use**</tt> on every modification. Accordingly getters must strip the
3153 For layout b) instead of the <tt>User</tt> we find a pointer (<tt>User*</tt> with LSBit set).
3154 Following this pointer brings us to the <tt>User</tt>. A portable trick ensures
3155 that the first bytes of <tt>User</tt> (if interpreted as a pointer) never has
3156 the LSBit set. (Portability is relying on the fact that all known compilers place the
3157 <tt>vptr</tt> in the first word of the instances.)</p>
3165 <!-- *********************************************************************** -->
3167 <a name="coreclasses">The Core LLVM Class Hierarchy Reference </a>
3169 <!-- *********************************************************************** -->
3172 <p><tt>#include "<a href="/doxygen/Type_8h-source.html">llvm/Type.h</a>"</tt>
3173 <br>doxygen info: <a href="/doxygen/classllvm_1_1Type.html">Type Class</a></p>
3175 <p>The Core LLVM classes are the primary means of representing the program
3176 being inspected or transformed. The core LLVM classes are defined in
3177 header files in the <tt>include/llvm/</tt> directory, and implemented in
3178 the <tt>lib/VMCore</tt> directory.</p>
3180 <!-- ======================================================================= -->
3182 <a name="Type">The <tt>Type</tt> class and Derived Types</a>
3187 <p><tt>Type</tt> is a superclass of all type classes. Every <tt>Value</tt> has
3188 a <tt>Type</tt>. <tt>Type</tt> cannot be instantiated directly but only
3189 through its subclasses. Certain primitive types (<tt>VoidType</tt>,
3190 <tt>LabelType</tt>, <tt>FloatType</tt> and <tt>DoubleType</tt>) have hidden
3191 subclasses. They are hidden because they offer no useful functionality beyond
3192 what the <tt>Type</tt> class offers except to distinguish themselves from
3193 other subclasses of <tt>Type</tt>.</p>
3194 <p>All other types are subclasses of <tt>DerivedType</tt>. Types can be
3195 named, but this is not a requirement. There exists exactly
3196 one instance of a given shape at any one time. This allows type equality to
3197 be performed with address equality of the Type Instance. That is, given two
3198 <tt>Type*</tt> values, the types are identical if the pointers are identical.
3201 <!-- _______________________________________________________________________ -->
3203 <a name="m_Type">Important Public Methods</a>
3209 <li><tt>bool isIntegerTy() const</tt>: Returns true for any integer type.</li>
3211 <li><tt>bool isFloatingPointTy()</tt>: Return true if this is one of the five
3212 floating point types.</li>
3214 <li><tt>bool isSized()</tt>: Return true if the type has known size. Things
3215 that don't have a size are abstract types, labels and void.</li>
3220 <!-- _______________________________________________________________________ -->
3222 <a name="derivedtypes">Important Derived Types</a>
3226 <dt><tt>IntegerType</tt></dt>
3227 <dd>Subclass of DerivedType that represents integer types of any bit width.
3228 Any bit width between <tt>IntegerType::MIN_INT_BITS</tt> (1) and
3229 <tt>IntegerType::MAX_INT_BITS</tt> (~8 million) can be represented.
3231 <li><tt>static const IntegerType* get(unsigned NumBits)</tt>: get an integer
3232 type of a specific bit width.</li>
3233 <li><tt>unsigned getBitWidth() const</tt>: Get the bit width of an integer
3237 <dt><tt>SequentialType</tt></dt>
3238 <dd>This is subclassed by ArrayType, PointerType and VectorType.
3240 <li><tt>const Type * getElementType() const</tt>: Returns the type of each
3241 of the elements in the sequential type. </li>
3244 <dt><tt>ArrayType</tt></dt>
3245 <dd>This is a subclass of SequentialType and defines the interface for array
3248 <li><tt>unsigned getNumElements() const</tt>: Returns the number of
3249 elements in the array. </li>
3252 <dt><tt>PointerType</tt></dt>
3253 <dd>Subclass of SequentialType for pointer types.</dd>
3254 <dt><tt>VectorType</tt></dt>
3255 <dd>Subclass of SequentialType for vector types. A
3256 vector type is similar to an ArrayType but is distinguished because it is
3257 a first class type whereas ArrayType is not. Vector types are used for
3258 vector operations and are usually small vectors of of an integer or floating
3260 <dt><tt>StructType</tt></dt>
3261 <dd>Subclass of DerivedTypes for struct types.</dd>
3262 <dt><tt><a name="FunctionType">FunctionType</a></tt></dt>
3263 <dd>Subclass of DerivedTypes for function types.
3265 <li><tt>bool isVarArg() const</tt>: Returns true if it's a vararg
3267 <li><tt> const Type * getReturnType() const</tt>: Returns the
3268 return type of the function.</li>
3269 <li><tt>const Type * getParamType (unsigned i)</tt>: Returns
3270 the type of the ith parameter.</li>
3271 <li><tt> const unsigned getNumParams() const</tt>: Returns the
3272 number of formal parameters.</li>
3280 <!-- ======================================================================= -->
3282 <a name="Module">The <tt>Module</tt> class</a>
3288 href="/doxygen/Module_8h-source.html">llvm/Module.h</a>"</tt><br> doxygen info:
3289 <a href="/doxygen/classllvm_1_1Module.html">Module Class</a></p>
3291 <p>The <tt>Module</tt> class represents the top level structure present in LLVM
3292 programs. An LLVM module is effectively either a translation unit of the
3293 original program or a combination of several translation units merged by the
3294 linker. The <tt>Module</tt> class keeps track of a list of <a
3295 href="#Function"><tt>Function</tt></a>s, a list of <a
3296 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s, and a <a
3297 href="#SymbolTable"><tt>SymbolTable</tt></a>. Additionally, it contains a few
3298 helpful member functions that try to make common operations easy.</p>
3300 <!-- _______________________________________________________________________ -->
3302 <a name="m_Module">Important Public Members of the <tt>Module</tt> class</a>
3308 <li><tt>Module::Module(std::string name = "")</tt></li>
3311 <p>Constructing a <a href="#Module">Module</a> is easy. You can optionally
3312 provide a name for it (probably based on the name of the translation unit).</p>
3315 <li><tt>Module::iterator</tt> - Typedef for function list iterator<br>
3316 <tt>Module::const_iterator</tt> - Typedef for const_iterator.<br>
3318 <tt>begin()</tt>, <tt>end()</tt>
3319 <tt>size()</tt>, <tt>empty()</tt>
3321 <p>These are forwarding methods that make it easy to access the contents of
3322 a <tt>Module</tt> object's <a href="#Function"><tt>Function</tt></a>
3325 <li><tt>Module::FunctionListType &getFunctionList()</tt>
3327 <p> Returns the list of <a href="#Function"><tt>Function</tt></a>s. This is
3328 necessary to use when you need to update the list or perform a complex
3329 action that doesn't have a forwarding method.</p>
3331 <p><!-- Global Variable --></p></li>
3337 <li><tt>Module::global_iterator</tt> - Typedef for global variable list iterator<br>
3339 <tt>Module::const_global_iterator</tt> - Typedef for const_iterator.<br>
3341 <tt>global_begin()</tt>, <tt>global_end()</tt>
3342 <tt>global_size()</tt>, <tt>global_empty()</tt>
3344 <p> These are forwarding methods that make it easy to access the contents of
3345 a <tt>Module</tt> object's <a
3346 href="#GlobalVariable"><tt>GlobalVariable</tt></a> list.</p></li>
3348 <li><tt>Module::GlobalListType &getGlobalList()</tt>
3350 <p>Returns the list of <a
3351 href="#GlobalVariable"><tt>GlobalVariable</tt></a>s. This is necessary to
3352 use when you need to update the list or perform a complex action that
3353 doesn't have a forwarding method.</p>
3355 <p><!-- Symbol table stuff --> </p></li>
3361 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
3363 <p>Return a reference to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3364 for this <tt>Module</tt>.</p>
3366 <p><!-- Convenience methods --></p></li>
3372 <li><tt><a href="#Function">Function</a> *getFunction(const std::string
3373 &Name, const <a href="#FunctionType">FunctionType</a> *Ty)</tt>
3375 <p>Look up the specified function in the <tt>Module</tt> <a
3376 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, return
3377 <tt>null</tt>.</p></li>
3379 <li><tt><a href="#Function">Function</a> *getOrInsertFunction(const
3380 std::string &Name, const <a href="#FunctionType">FunctionType</a> *T)</tt>
3382 <p>Look up the specified function in the <tt>Module</tt> <a
3383 href="#SymbolTable"><tt>SymbolTable</tt></a>. If it does not exist, add an
3384 external declaration for the function and return it.</p></li>
3386 <li><tt>std::string getTypeName(const <a href="#Type">Type</a> *Ty)</tt>
3388 <p>If there is at least one entry in the <a
3389 href="#SymbolTable"><tt>SymbolTable</tt></a> for the specified <a
3390 href="#Type"><tt>Type</tt></a>, return it. Otherwise return the empty
3393 <li><tt>bool addTypeName(const std::string &Name, const <a
3394 href="#Type">Type</a> *Ty)</tt>
3396 <p>Insert an entry in the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3397 mapping <tt>Name</tt> to <tt>Ty</tt>. If there is already an entry for this
3398 name, true is returned and the <a
3399 href="#SymbolTable"><tt>SymbolTable</tt></a> is not modified.</p></li>
3406 <!-- ======================================================================= -->
3408 <a name="Value">The <tt>Value</tt> class</a>
3413 <p><tt>#include "<a href="/doxygen/Value_8h-source.html">llvm/Value.h</a>"</tt>
3415 doxygen info: <a href="/doxygen/classllvm_1_1Value.html">Value Class</a></p>
3417 <p>The <tt>Value</tt> class is the most important class in the LLVM Source
3418 base. It represents a typed value that may be used (among other things) as an
3419 operand to an instruction. There are many different types of <tt>Value</tt>s,
3420 such as <a href="#Constant"><tt>Constant</tt></a>s,<a
3421 href="#Argument"><tt>Argument</tt></a>s. Even <a
3422 href="#Instruction"><tt>Instruction</tt></a>s and <a
3423 href="#Function"><tt>Function</tt></a>s are <tt>Value</tt>s.</p>
3425 <p>A particular <tt>Value</tt> may be used many times in the LLVM representation
3426 for a program. For example, an incoming argument to a function (represented
3427 with an instance of the <a href="#Argument">Argument</a> class) is "used" by
3428 every instruction in the function that references the argument. To keep track
3429 of this relationship, the <tt>Value</tt> class keeps a list of all of the <a
3430 href="#User"><tt>User</tt></a>s that is using it (the <a
3431 href="#User"><tt>User</tt></a> class is a base class for all nodes in the LLVM
3432 graph that can refer to <tt>Value</tt>s). This use list is how LLVM represents
3433 def-use information in the program, and is accessible through the <tt>use_</tt>*
3434 methods, shown below.</p>
3436 <p>Because LLVM is a typed representation, every LLVM <tt>Value</tt> is typed,
3437 and this <a href="#Type">Type</a> is available through the <tt>getType()</tt>
3438 method. In addition, all LLVM values can be named. The "name" of the
3439 <tt>Value</tt> is a symbolic string printed in the LLVM code:</p>
3441 <div class="doc_code">
3443 %<b>foo</b> = add i32 1, 2
3447 <p><a name="nameWarning">The name of this instruction is "foo".</a> <b>NOTE</b>
3448 that the name of any value may be missing (an empty string), so names should
3449 <b>ONLY</b> be used for debugging (making the source code easier to read,
3450 debugging printouts), they should not be used to keep track of values or map
3451 between them. For this purpose, use a <tt>std::map</tt> of pointers to the
3452 <tt>Value</tt> itself instead.</p>
3454 <p>One important aspect of LLVM is that there is no distinction between an SSA
3455 variable and the operation that produces it. Because of this, any reference to
3456 the value produced by an instruction (or the value available as an incoming
3457 argument, for example) is represented as a direct pointer to the instance of
3459 represents this value. Although this may take some getting used to, it
3460 simplifies the representation and makes it easier to manipulate.</p>
3462 <!-- _______________________________________________________________________ -->
3464 <a name="m_Value">Important Public Members of the <tt>Value</tt> class</a>
3470 <li><tt>Value::use_iterator</tt> - Typedef for iterator over the
3472 <tt>Value::const_use_iterator</tt> - Typedef for const_iterator over
3474 <tt>unsigned use_size()</tt> - Returns the number of users of the
3476 <tt>bool use_empty()</tt> - Returns true if there are no users.<br>
3477 <tt>use_iterator use_begin()</tt> - Get an iterator to the start of
3479 <tt>use_iterator use_end()</tt> - Get an iterator to the end of the
3481 <tt><a href="#User">User</a> *use_back()</tt> - Returns the last
3482 element in the list.
3483 <p> These methods are the interface to access the def-use
3484 information in LLVM. As with all other iterators in LLVM, the naming
3485 conventions follow the conventions defined by the <a href="#stl">STL</a>.</p>
3487 <li><tt><a href="#Type">Type</a> *getType() const</tt>
3488 <p>This method returns the Type of the Value.</p>
3490 <li><tt>bool hasName() const</tt><br>
3491 <tt>std::string getName() const</tt><br>
3492 <tt>void setName(const std::string &Name)</tt>
3493 <p> This family of methods is used to access and assign a name to a <tt>Value</tt>,
3494 be aware of the <a href="#nameWarning">precaution above</a>.</p>
3496 <li><tt>void replaceAllUsesWith(Value *V)</tt>
3498 <p>This method traverses the use list of a <tt>Value</tt> changing all <a
3499 href="#User"><tt>User</tt>s</a> of the current value to refer to
3500 "<tt>V</tt>" instead. For example, if you detect that an instruction always
3501 produces a constant value (for example through constant folding), you can
3502 replace all uses of the instruction with the constant like this:</p>
3504 <div class="doc_code">
3506 Inst->replaceAllUsesWith(ConstVal);
3516 <!-- ======================================================================= -->
3518 <a name="User">The <tt>User</tt> class</a>
3524 <tt>#include "<a href="/doxygen/User_8h-source.html">llvm/User.h</a>"</tt><br>
3525 doxygen info: <a href="/doxygen/classllvm_1_1User.html">User Class</a><br>
3526 Superclass: <a href="#Value"><tt>Value</tt></a></p>
3528 <p>The <tt>User</tt> class is the common base class of all LLVM nodes that may
3529 refer to <a href="#Value"><tt>Value</tt></a>s. It exposes a list of "Operands"
3530 that are all of the <a href="#Value"><tt>Value</tt></a>s that the User is
3531 referring to. The <tt>User</tt> class itself is a subclass of
3534 <p>The operands of a <tt>User</tt> point directly to the LLVM <a
3535 href="#Value"><tt>Value</tt></a> that it refers to. Because LLVM uses Static
3536 Single Assignment (SSA) form, there can only be one definition referred to,
3537 allowing this direct connection. This connection provides the use-def
3538 information in LLVM.</p>
3540 <!-- _______________________________________________________________________ -->
3542 <a name="m_User">Important Public Members of the <tt>User</tt> class</a>
3547 <p>The <tt>User</tt> class exposes the operand list in two ways: through
3548 an index access interface and through an iterator based interface.</p>
3551 <li><tt>Value *getOperand(unsigned i)</tt><br>
3552 <tt>unsigned getNumOperands()</tt>
3553 <p> These two methods expose the operands of the <tt>User</tt> in a
3554 convenient form for direct access.</p></li>
3556 <li><tt>User::op_iterator</tt> - Typedef for iterator over the operand
3558 <tt>op_iterator op_begin()</tt> - Get an iterator to the start of
3559 the operand list.<br>
3560 <tt>op_iterator op_end()</tt> - Get an iterator to the end of the
3562 <p> Together, these methods make up the iterator based interface to
3563 the operands of a <tt>User</tt>.</p></li>
3570 <!-- ======================================================================= -->
3572 <a name="Instruction">The <tt>Instruction</tt> class</a>
3577 <p><tt>#include "</tt><tt><a
3578 href="/doxygen/Instruction_8h-source.html">llvm/Instruction.h</a>"</tt><br>
3579 doxygen info: <a href="/doxygen/classllvm_1_1Instruction.html">Instruction Class</a><br>
3580 Superclasses: <a href="#User"><tt>User</tt></a>, <a
3581 href="#Value"><tt>Value</tt></a></p>
3583 <p>The <tt>Instruction</tt> class is the common base class for all LLVM
3584 instructions. It provides only a few methods, but is a very commonly used
3585 class. The primary data tracked by the <tt>Instruction</tt> class itself is the
3586 opcode (instruction type) and the parent <a
3587 href="#BasicBlock"><tt>BasicBlock</tt></a> the <tt>Instruction</tt> is embedded
3588 into. To represent a specific type of instruction, one of many subclasses of
3589 <tt>Instruction</tt> are used.</p>
3591 <p> Because the <tt>Instruction</tt> class subclasses the <a
3592 href="#User"><tt>User</tt></a> class, its operands can be accessed in the same
3593 way as for other <a href="#User"><tt>User</tt></a>s (with the
3594 <tt>getOperand()</tt>/<tt>getNumOperands()</tt> and
3595 <tt>op_begin()</tt>/<tt>op_end()</tt> methods).</p> <p> An important file for
3596 the <tt>Instruction</tt> class is the <tt>llvm/Instruction.def</tt> file. This
3597 file contains some meta-data about the various different types of instructions
3598 in LLVM. It describes the enum values that are used as opcodes (for example
3599 <tt>Instruction::Add</tt> and <tt>Instruction::ICmp</tt>), as well as the
3600 concrete sub-classes of <tt>Instruction</tt> that implement the instruction (for
3601 example <tt><a href="#BinaryOperator">BinaryOperator</a></tt> and <tt><a
3602 href="#CmpInst">CmpInst</a></tt>). Unfortunately, the use of macros in
3603 this file confuses doxygen, so these enum values don't show up correctly in the
3604 <a href="/doxygen/classllvm_1_1Instruction.html">doxygen output</a>.</p>
3606 <!-- _______________________________________________________________________ -->
3608 <a name="s_Instruction">
3609 Important Subclasses of the <tt>Instruction</tt> class
3614 <li><tt><a name="BinaryOperator">BinaryOperator</a></tt>
3615 <p>This subclasses represents all two operand instructions whose operands
3616 must be the same type, except for the comparison instructions.</p></li>
3617 <li><tt><a name="CastInst">CastInst</a></tt>
3618 <p>This subclass is the parent of the 12 casting instructions. It provides
3619 common operations on cast instructions.</p>
3620 <li><tt><a name="CmpInst">CmpInst</a></tt>
3621 <p>This subclass respresents the two comparison instructions,
3622 <a href="LangRef.html#i_icmp">ICmpInst</a> (integer opreands), and
3623 <a href="LangRef.html#i_fcmp">FCmpInst</a> (floating point operands).</p>
3624 <li><tt><a name="TerminatorInst">TerminatorInst</a></tt>
3625 <p>This subclass is the parent of all terminator instructions (those which
3626 can terminate a block).</p>
3630 <!-- _______________________________________________________________________ -->
3632 <a name="m_Instruction">
3633 Important Public Members of the <tt>Instruction</tt> class
3640 <li><tt><a href="#BasicBlock">BasicBlock</a> *getParent()</tt>
3641 <p>Returns the <a href="#BasicBlock"><tt>BasicBlock</tt></a> that
3642 this <tt>Instruction</tt> is embedded into.</p></li>
3643 <li><tt>bool mayWriteToMemory()</tt>
3644 <p>Returns true if the instruction writes to memory, i.e. it is a
3645 <tt>call</tt>,<tt>free</tt>,<tt>invoke</tt>, or <tt>store</tt>.</p></li>
3646 <li><tt>unsigned getOpcode()</tt>
3647 <p>Returns the opcode for the <tt>Instruction</tt>.</p></li>
3648 <li><tt><a href="#Instruction">Instruction</a> *clone() const</tt>
3649 <p>Returns another instance of the specified instruction, identical
3650 in all ways to the original except that the instruction has no parent
3651 (ie it's not embedded into a <a href="#BasicBlock"><tt>BasicBlock</tt></a>),
3652 and it has no name</p></li>
3659 <!-- ======================================================================= -->
3661 <a name="Constant">The <tt>Constant</tt> class and subclasses</a>
3666 <p>Constant represents a base class for different types of constants. It
3667 is subclassed by ConstantInt, ConstantArray, etc. for representing
3668 the various types of Constants. <a href="#GlobalValue">GlobalValue</a> is also
3669 a subclass, which represents the address of a global variable or function.
3672 <!-- _______________________________________________________________________ -->
3673 <h4>Important Subclasses of Constant</h4>
3676 <li>ConstantInt : This subclass of Constant represents an integer constant of
3679 <li><tt>const APInt& getValue() const</tt>: Returns the underlying
3680 value of this constant, an APInt value.</li>
3681 <li><tt>int64_t getSExtValue() const</tt>: Converts the underlying APInt
3682 value to an int64_t via sign extension. If the value (not the bit width)
3683 of the APInt is too large to fit in an int64_t, an assertion will result.
3684 For this reason, use of this method is discouraged.</li>
3685 <li><tt>uint64_t getZExtValue() const</tt>: Converts the underlying APInt
3686 value to a uint64_t via zero extension. IF the value (not the bit width)
3687 of the APInt is too large to fit in a uint64_t, an assertion will result.
3688 For this reason, use of this method is discouraged.</li>
3689 <li><tt>static ConstantInt* get(const APInt& Val)</tt>: Returns the
3690 ConstantInt object that represents the value provided by <tt>Val</tt>.
3691 The type is implied as the IntegerType that corresponds to the bit width
3692 of <tt>Val</tt>.</li>
3693 <li><tt>static ConstantInt* get(const Type *Ty, uint64_t Val)</tt>:
3694 Returns the ConstantInt object that represents the value provided by
3695 <tt>Val</tt> for integer type <tt>Ty</tt>.</li>
3698 <li>ConstantFP : This class represents a floating point constant.
3700 <li><tt>double getValue() const</tt>: Returns the underlying value of
3701 this constant. </li>
3704 <li>ConstantArray : This represents a constant array.
3706 <li><tt>const std::vector<Use> &getValues() const</tt>: Returns
3707 a vector of component constants that makeup this array. </li>
3710 <li>ConstantStruct : This represents a constant struct.
3712 <li><tt>const std::vector<Use> &getValues() const</tt>: Returns
3713 a vector of component constants that makeup this array. </li>
3716 <li>GlobalValue : This represents either a global variable or a function. In
3717 either case, the value is a constant fixed address (after linking).
3724 <!-- ======================================================================= -->
3726 <a name="GlobalValue">The <tt>GlobalValue</tt> class</a>
3732 href="/doxygen/GlobalValue_8h-source.html">llvm/GlobalValue.h</a>"</tt><br>
3733 doxygen info: <a href="/doxygen/classllvm_1_1GlobalValue.html">GlobalValue
3735 Superclasses: <a href="#Constant"><tt>Constant</tt></a>,
3736 <a href="#User"><tt>User</tt></a>, <a href="#Value"><tt>Value</tt></a></p>
3738 <p>Global values (<a href="#GlobalVariable"><tt>GlobalVariable</tt></a>s or <a
3739 href="#Function"><tt>Function</tt></a>s) are the only LLVM values that are
3740 visible in the bodies of all <a href="#Function"><tt>Function</tt></a>s.
3741 Because they are visible at global scope, they are also subject to linking with
3742 other globals defined in different translation units. To control the linking
3743 process, <tt>GlobalValue</tt>s know their linkage rules. Specifically,
3744 <tt>GlobalValue</tt>s know whether they have internal or external linkage, as
3745 defined by the <tt>LinkageTypes</tt> enumeration.</p>
3747 <p>If a <tt>GlobalValue</tt> has internal linkage (equivalent to being
3748 <tt>static</tt> in C), it is not visible to code outside the current translation
3749 unit, and does not participate in linking. If it has external linkage, it is
3750 visible to external code, and does participate in linking. In addition to
3751 linkage information, <tt>GlobalValue</tt>s keep track of which <a
3752 href="#Module"><tt>Module</tt></a> they are currently part of.</p>
3754 <p>Because <tt>GlobalValue</tt>s are memory objects, they are always referred to
3755 by their <b>address</b>. As such, the <a href="#Type"><tt>Type</tt></a> of a
3756 global is always a pointer to its contents. It is important to remember this
3757 when using the <tt>GetElementPtrInst</tt> instruction because this pointer must
3758 be dereferenced first. For example, if you have a <tt>GlobalVariable</tt> (a
3759 subclass of <tt>GlobalValue)</tt> that is an array of 24 ints, type <tt>[24 x
3760 i32]</tt>, then the <tt>GlobalVariable</tt> is a pointer to that array. Although
3761 the address of the first element of this array and the value of the
3762 <tt>GlobalVariable</tt> are the same, they have different types. The
3763 <tt>GlobalVariable</tt>'s type is <tt>[24 x i32]</tt>. The first element's type
3764 is <tt>i32.</tt> Because of this, accessing a global value requires you to
3765 dereference the pointer with <tt>GetElementPtrInst</tt> first, then its elements
3766 can be accessed. This is explained in the <a href="LangRef.html#globalvars">LLVM
3767 Language Reference Manual</a>.</p>
3769 <!-- _______________________________________________________________________ -->
3771 <a name="m_GlobalValue">
3772 Important Public Members of the <tt>GlobalValue</tt> class
3779 <li><tt>bool hasInternalLinkage() const</tt><br>
3780 <tt>bool hasExternalLinkage() const</tt><br>
3781 <tt>void setInternalLinkage(bool HasInternalLinkage)</tt>
3782 <p> These methods manipulate the linkage characteristics of the <tt>GlobalValue</tt>.</p>
3785 <li><tt><a href="#Module">Module</a> *getParent()</tt>
3786 <p> This returns the <a href="#Module"><tt>Module</tt></a> that the
3787 GlobalValue is currently embedded into.</p></li>
3794 <!-- ======================================================================= -->
3796 <a name="Function">The <tt>Function</tt> class</a>
3802 href="/doxygen/Function_8h-source.html">llvm/Function.h</a>"</tt><br> doxygen
3803 info: <a href="/doxygen/classllvm_1_1Function.html">Function Class</a><br>
3804 Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3805 <a href="#Constant"><tt>Constant</tt></a>,
3806 <a href="#User"><tt>User</tt></a>,
3807 <a href="#Value"><tt>Value</tt></a></p>
3809 <p>The <tt>Function</tt> class represents a single procedure in LLVM. It is
3810 actually one of the more complex classes in the LLVM hierarchy because it must
3811 keep track of a large amount of data. The <tt>Function</tt> class keeps track
3812 of a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, a list of formal
3813 <a href="#Argument"><tt>Argument</tt></a>s, and a
3814 <a href="#SymbolTable"><tt>SymbolTable</tt></a>.</p>
3816 <p>The list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s is the most
3817 commonly used part of <tt>Function</tt> objects. The list imposes an implicit
3818 ordering of the blocks in the function, which indicate how the code will be
3819 laid out by the backend. Additionally, the first <a
3820 href="#BasicBlock"><tt>BasicBlock</tt></a> is the implicit entry node for the
3821 <tt>Function</tt>. It is not legal in LLVM to explicitly branch to this initial
3822 block. There are no implicit exit nodes, and in fact there may be multiple exit
3823 nodes from a single <tt>Function</tt>. If the <a
3824 href="#BasicBlock"><tt>BasicBlock</tt></a> list is empty, this indicates that
3825 the <tt>Function</tt> is actually a function declaration: the actual body of the
3826 function hasn't been linked in yet.</p>
3828 <p>In addition to a list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s, the
3829 <tt>Function</tt> class also keeps track of the list of formal <a
3830 href="#Argument"><tt>Argument</tt></a>s that the function receives. This
3831 container manages the lifetime of the <a href="#Argument"><tt>Argument</tt></a>
3832 nodes, just like the <a href="#BasicBlock"><tt>BasicBlock</tt></a> list does for
3833 the <a href="#BasicBlock"><tt>BasicBlock</tt></a>s.</p>
3835 <p>The <a href="#SymbolTable"><tt>SymbolTable</tt></a> is a very rarely used
3836 LLVM feature that is only used when you have to look up a value by name. Aside
3837 from that, the <a href="#SymbolTable"><tt>SymbolTable</tt></a> is used
3838 internally to make sure that there are not conflicts between the names of <a
3839 href="#Instruction"><tt>Instruction</tt></a>s, <a
3840 href="#BasicBlock"><tt>BasicBlock</tt></a>s, or <a
3841 href="#Argument"><tt>Argument</tt></a>s in the function body.</p>
3843 <p>Note that <tt>Function</tt> is a <a href="#GlobalValue">GlobalValue</a>
3844 and therefore also a <a href="#Constant">Constant</a>. The value of the function
3845 is its address (after linking) which is guaranteed to be constant.</p>
3847 <!-- _______________________________________________________________________ -->
3849 <a name="m_Function">
3850 Important Public Members of the <tt>Function</tt> class
3857 <li><tt>Function(const </tt><tt><a href="#FunctionType">FunctionType</a>
3858 *Ty, LinkageTypes Linkage, const std::string &N = "", Module* Parent = 0)</tt>
3860 <p>Constructor used when you need to create new <tt>Function</tt>s to add
3861 the the program. The constructor must specify the type of the function to
3862 create and what type of linkage the function should have. The <a
3863 href="#FunctionType"><tt>FunctionType</tt></a> argument
3864 specifies the formal arguments and return value for the function. The same
3865 <a href="#FunctionType"><tt>FunctionType</tt></a> value can be used to
3866 create multiple functions. The <tt>Parent</tt> argument specifies the Module
3867 in which the function is defined. If this argument is provided, the function
3868 will automatically be inserted into that module's list of
3871 <li><tt>bool isDeclaration()</tt>
3873 <p>Return whether or not the <tt>Function</tt> has a body defined. If the
3874 function is "external", it does not have a body, and thus must be resolved
3875 by linking with a function defined in a different translation unit.</p></li>
3877 <li><tt>Function::iterator</tt> - Typedef for basic block list iterator<br>
3878 <tt>Function::const_iterator</tt> - Typedef for const_iterator.<br>
3880 <tt>begin()</tt>, <tt>end()</tt>
3881 <tt>size()</tt>, <tt>empty()</tt>
3883 <p>These are forwarding methods that make it easy to access the contents of
3884 a <tt>Function</tt> object's <a href="#BasicBlock"><tt>BasicBlock</tt></a>
3887 <li><tt>Function::BasicBlockListType &getBasicBlockList()</tt>
3889 <p>Returns the list of <a href="#BasicBlock"><tt>BasicBlock</tt></a>s. This
3890 is necessary to use when you need to update the list or perform a complex
3891 action that doesn't have a forwarding method.</p></li>
3893 <li><tt>Function::arg_iterator</tt> - Typedef for the argument list
3895 <tt>Function::const_arg_iterator</tt> - Typedef for const_iterator.<br>
3897 <tt>arg_begin()</tt>, <tt>arg_end()</tt>
3898 <tt>arg_size()</tt>, <tt>arg_empty()</tt>
3900 <p>These are forwarding methods that make it easy to access the contents of
3901 a <tt>Function</tt> object's <a href="#Argument"><tt>Argument</tt></a>
3904 <li><tt>Function::ArgumentListType &getArgumentList()</tt>
3906 <p>Returns the list of <a href="#Argument"><tt>Argument</tt></a>s. This is
3907 necessary to use when you need to update the list or perform a complex
3908 action that doesn't have a forwarding method.</p></li>
3910 <li><tt><a href="#BasicBlock">BasicBlock</a> &getEntryBlock()</tt>
3912 <p>Returns the entry <a href="#BasicBlock"><tt>BasicBlock</tt></a> for the
3913 function. Because the entry block for the function is always the first
3914 block, this returns the first block of the <tt>Function</tt>.</p></li>
3916 <li><tt><a href="#Type">Type</a> *getReturnType()</tt><br>
3917 <tt><a href="#FunctionType">FunctionType</a> *getFunctionType()</tt>
3919 <p>This traverses the <a href="#Type"><tt>Type</tt></a> of the
3920 <tt>Function</tt> and returns the return type of the function, or the <a
3921 href="#FunctionType"><tt>FunctionType</tt></a> of the actual
3924 <li><tt><a href="#SymbolTable">SymbolTable</a> *getSymbolTable()</tt>
3926 <p> Return a pointer to the <a href="#SymbolTable"><tt>SymbolTable</tt></a>
3927 for this <tt>Function</tt>.</p></li>
3934 <!-- ======================================================================= -->
3936 <a name="GlobalVariable">The <tt>GlobalVariable</tt> class</a>
3942 href="/doxygen/GlobalVariable_8h-source.html">llvm/GlobalVariable.h</a>"</tt>
3944 doxygen info: <a href="/doxygen/classllvm_1_1GlobalVariable.html">GlobalVariable
3946 Superclasses: <a href="#GlobalValue"><tt>GlobalValue</tt></a>,
3947 <a href="#Constant"><tt>Constant</tt></a>,
3948 <a href="#User"><tt>User</tt></a>,
3949 <a href="#Value"><tt>Value</tt></a></p>
3951 <p>Global variables are represented with the (surprise surprise)
3952 <tt>GlobalVariable</tt> class. Like functions, <tt>GlobalVariable</tt>s are also
3953 subclasses of <a href="#GlobalValue"><tt>GlobalValue</tt></a>, and as such are
3954 always referenced by their address (global values must live in memory, so their
3955 "name" refers to their constant address). See
3956 <a href="#GlobalValue"><tt>GlobalValue</tt></a> for more on this. Global
3957 variables may have an initial value (which must be a
3958 <a href="#Constant"><tt>Constant</tt></a>), and if they have an initializer,
3959 they may be marked as "constant" themselves (indicating that their contents
3960 never change at runtime).</p>
3962 <!-- _______________________________________________________________________ -->
3964 <a name="m_GlobalVariable">
3965 Important Public Members of the <tt>GlobalVariable</tt> class
3972 <li><tt>GlobalVariable(const </tt><tt><a href="#Type">Type</a> *Ty, bool
3973 isConstant, LinkageTypes& Linkage, <a href="#Constant">Constant</a>
3974 *Initializer = 0, const std::string &Name = "", Module* Parent = 0)</tt>
3976 <p>Create a new global variable of the specified type. If
3977 <tt>isConstant</tt> is true then the global variable will be marked as
3978 unchanging for the program. The Linkage parameter specifies the type of
3979 linkage (internal, external, weak, linkonce, appending) for the variable.
3980 If the linkage is InternalLinkage, WeakAnyLinkage, WeakODRLinkage,
3981 LinkOnceAnyLinkage or LinkOnceODRLinkage, then the resultant
3982 global variable will have internal linkage. AppendingLinkage concatenates
3983 together all instances (in different translation units) of the variable
3984 into a single variable but is only applicable to arrays. See
3985 the <a href="LangRef.html#modulestructure">LLVM Language Reference</a> for
3986 further details on linkage types. Optionally an initializer, a name, and the
3987 module to put the variable into may be specified for the global variable as
3990 <li><tt>bool isConstant() const</tt>
3992 <p>Returns true if this is a global variable that is known not to
3993 be modified at runtime.</p></li>
3995 <li><tt>bool hasInitializer()</tt>
3997 <p>Returns true if this <tt>GlobalVariable</tt> has an intializer.</p></li>
3999 <li><tt><a href="#Constant">Constant</a> *getInitializer()</tt>
4001 <p>Returns the initial value for a <tt>GlobalVariable</tt>. It is not legal
4002 to call this method if there is no initializer.</p></li>
4009 <!-- ======================================================================= -->
4011 <a name="BasicBlock">The <tt>BasicBlock</tt> class</a>
4017 href="/doxygen/BasicBlock_8h-source.html">llvm/BasicBlock.h</a>"</tt><br>
4018 doxygen info: <a href="/doxygen/classllvm_1_1BasicBlock.html">BasicBlock
4020 Superclass: <a href="#Value"><tt>Value</tt></a></p>
4022 <p>This class represents a single entry single exit section of the code,
4023 commonly known as a basic block by the compiler community. The
4024 <tt>BasicBlock</tt> class maintains a list of <a
4025 href="#Instruction"><tt>Instruction</tt></a>s, which form the body of the block.
4026 Matching the language definition, the last element of this list of instructions
4027 is always a terminator instruction (a subclass of the <a
4028 href="#TerminatorInst"><tt>TerminatorInst</tt></a> class).</p>
4030 <p>In addition to tracking the list of instructions that make up the block, the
4031 <tt>BasicBlock</tt> class also keeps track of the <a
4032 href="#Function"><tt>Function</tt></a> that it is embedded into.</p>
4034 <p>Note that <tt>BasicBlock</tt>s themselves are <a
4035 href="#Value"><tt>Value</tt></a>s, because they are referenced by instructions
4036 like branches and can go in the switch tables. <tt>BasicBlock</tt>s have type
4039 <!-- _______________________________________________________________________ -->
4041 <a name="m_BasicBlock">
4042 Important Public Members of the <tt>BasicBlock</tt> class
4049 <li><tt>BasicBlock(const std::string &Name = "", </tt><tt><a
4050 href="#Function">Function</a> *Parent = 0)</tt>
4052 <p>The <tt>BasicBlock</tt> constructor is used to create new basic blocks for
4053 insertion into a function. The constructor optionally takes a name for the new
4054 block, and a <a href="#Function"><tt>Function</tt></a> to insert it into. If
4055 the <tt>Parent</tt> parameter is specified, the new <tt>BasicBlock</tt> is
4056 automatically inserted at the end of the specified <a
4057 href="#Function"><tt>Function</tt></a>, if not specified, the BasicBlock must be
4058 manually inserted into the <a href="#Function"><tt>Function</tt></a>.</p></li>
4060 <li><tt>BasicBlock::iterator</tt> - Typedef for instruction list iterator<br>
4061 <tt>BasicBlock::const_iterator</tt> - Typedef for const_iterator.<br>
4062 <tt>begin()</tt>, <tt>end()</tt>, <tt>front()</tt>, <tt>back()</tt>,
4063 <tt>size()</tt>, <tt>empty()</tt>
4064 STL-style functions for accessing the instruction list.
4066 <p>These methods and typedefs are forwarding functions that have the same
4067 semantics as the standard library methods of the same names. These methods
4068 expose the underlying instruction list of a basic block in a way that is easy to
4069 manipulate. To get the full complement of container operations (including
4070 operations to update the list), you must use the <tt>getInstList()</tt>
4073 <li><tt>BasicBlock::InstListType &getInstList()</tt>
4075 <p>This method is used to get access to the underlying container that actually
4076 holds the Instructions. This method must be used when there isn't a forwarding
4077 function in the <tt>BasicBlock</tt> class for the operation that you would like
4078 to perform. Because there are no forwarding functions for "updating"
4079 operations, you need to use this if you want to update the contents of a
4080 <tt>BasicBlock</tt>.</p></li>
4082 <li><tt><a href="#Function">Function</a> *getParent()</tt>
4084 <p> Returns a pointer to <a href="#Function"><tt>Function</tt></a> the block is
4085 embedded into, or a null pointer if it is homeless.</p></li>
4087 <li><tt><a href="#TerminatorInst">TerminatorInst</a> *getTerminator()</tt>
4089 <p> Returns a pointer to the terminator instruction that appears at the end of
4090 the <tt>BasicBlock</tt>. If there is no terminator instruction, or if the last
4091 instruction in the block is not a terminator, then a null pointer is
4100 <!-- ======================================================================= -->
4102 <a name="Argument">The <tt>Argument</tt> class</a>
4107 <p>This subclass of Value defines the interface for incoming formal
4108 arguments to a function. A Function maintains a list of its formal
4109 arguments. An argument has a pointer to the parent Function.</p>
4115 <!-- *********************************************************************** -->
4118 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
4119 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
4120 <a href="http://validator.w3.org/check/referer"><img
4121 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01 Strict"></a>
4123 <a href="mailto:dhurjati@cs.uiuc.edu">Dinakar Dhurjati</a> and
4124 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
4125 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
4126 Last modified: $Date$