blob: 2a1d3a1bda192773d2f9fedafe149538e298b6dd [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman9d0919f2003-11-08 01:05:38 +00003<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +00006 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
8 <meta name="description"
9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman9d0919f2003-11-08 01:05:38 +000010 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +000012
Misha Brukman9d0919f2003-11-08 01:05:38 +000013<body>
Chris Lattnerd7923912004-05-23 21:06:01 +000014
Chris Lattner261efe92003-11-25 01:02:51 +000015<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner00950542001-06-06 20:29:01 +000016<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +000017 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000020 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Chris Lattnere5d947b2004-12-09 16:36:40 +000023 <li><a href="#linkage">Linkage Types</a></li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +000024 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000025 <li><a href="#globalvars">Global Variables</a></li>
26 <li><a href="#functionstructure">Function Structure</a></li>
27 </ol>
28 </li>
Chris Lattner00950542001-06-06 20:29:01 +000029 <li><a href="#typesystem">Type System</a>
30 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000031 <li><a href="#t_primitive">Primitive Types</a>
32 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +000033 <li><a href="#t_classifications">Type Classifications</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000034 </ol>
35 </li>
Chris Lattner00950542001-06-06 20:29:01 +000036 <li><a href="#t_derived">Derived Types</a>
37 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000038 <li><a href="#t_array">Array Type</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000039 <li><a href="#t_function">Function Type</a></li>
40 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000041 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattnera58561b2004-08-12 19:12:28 +000042 <li><a href="#t_packed">Packed Type</a></li>
Chris Lattner69c11bb2005-04-25 17:34:15 +000043 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000044 </ol>
45 </li>
46 </ol>
47 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +000048 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +000049 <ol>
50 <li><a href="#simpleconstants">Simple Constants</a>
51 <li><a href="#aggregateconstants">Aggregate Constants</a>
52 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
53 <li><a href="#undefvalues">Undefined Values</a>
54 <li><a href="#constantexprs">Constant Expressions</a>
55 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +000056 </li>
Chris Lattner00950542001-06-06 20:29:01 +000057 <li><a href="#instref">Instruction Reference</a>
58 <ol>
59 <li><a href="#terminators">Terminator Instructions</a>
60 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000061 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
62 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000063 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
64 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000065 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +000066 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000067 </ol>
68 </li>
Chris Lattner00950542001-06-06 20:29:01 +000069 <li><a href="#binaryops">Binary Operations</a>
70 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000071 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
72 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
73 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
74 <li><a href="#i_div">'<tt>div</tt>' Instruction</a></li>
75 <li><a href="#i_rem">'<tt>rem</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000076 <li><a href="#i_setcc">'<tt>set<i>cc</i></tt>' Instructions</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000077 </ol>
78 </li>
Chris Lattner00950542001-06-06 20:29:01 +000079 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
80 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +000081 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000082 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000083 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
84 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
85 <li><a href="#i_shr">'<tt>shr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000086 </ol>
87 </li>
Chris Lattner00950542001-06-06 20:29:01 +000088 <li><a href="#memoryops">Memory Access Operations</a>
89 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000090 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
91 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
92 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
93 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
94 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
95 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
96 </ol>
97 </li>
Chris Lattner00950542001-06-06 20:29:01 +000098 <li><a href="#otherops">Other Operations</a>
99 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000100 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000101 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000102 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Robert Bocchino3a558662006-01-05 17:37:02 +0000103 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000104 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000105 <li><a href="#i_vaarg">'<tt>vaarg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000106 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000107 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000108 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000109 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000110 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000111 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000112 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
113 <ol>
114 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
115 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
116 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
117 </ol>
118 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000119 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
120 <ol>
121 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
122 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
123 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
124 </ol>
125 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000126 <li><a href="#int_codegen">Code Generator Intrinsics</a>
127 <ol>
128 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
129 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000130 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000131 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000132 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000133 </ol>
134 </li>
135 <li><a href="#int_os">Operating System Intrinsics</a>
136 <ol>
Chris Lattner32006282004-06-11 02:28:03 +0000137 <li><a href="#i_readport">'<tt>llvm.readport</tt>' Intrinsic</a></li>
138 <li><a href="#i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a></li>
John Criswell183402a2004-04-12 15:02:16 +0000139 <li><a href="#i_readio">'<tt>llvm.readio</tt>' Intrinsic</a></li>
140 <li><a href="#i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000141 </ol>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000142 <li><a href="#int_libc">Standard C Library Intrinsics</a>
143 <ol>
144 <li><a href="#i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a></li>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000145 <li><a href="#i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000146 <li><a href="#i_memset">'<tt>llvm.memset</tt>' Intrinsic</a></li>
Alkis Evlogimenos96853722004-06-12 19:19:14 +0000147 <li><a href="#i_isunordered">'<tt>llvm.isunordered</tt>' Intrinsic</a></li>
Chris Lattnera4d74142005-07-21 01:29:16 +0000148 <li><a href="#i_sqrt">'<tt>llvm.sqrt</tt>' Intrinsic</a></li>
149
Chris Lattner33aec9e2004-02-12 17:01:32 +0000150 </ol>
151 </li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000152 <li><a href="#int_count">Bit counting Intrinsics</a>
153 <ol>
154 <li><a href="#int_ctpop">'<tt>llvm.ctpop</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000155 <li><a href="#int_ctlz">'<tt>llvm.ctlz</tt>' Intrinsic </a></li>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000156 <li><a href="#int_cttz">'<tt>llvm.cttz</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000157 </ol>
158 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000159 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000160 </ol>
161 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000162</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000163
164<div class="doc_author">
165 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
166 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000167</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000168
Chris Lattner00950542001-06-06 20:29:01 +0000169<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000170<div class="doc_section"> <a name="abstract">Abstract </a></div>
171<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000172
Misha Brukman9d0919f2003-11-08 01:05:38 +0000173<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000174<p>This document is a reference manual for the LLVM assembly language.
175LLVM is an SSA based representation that provides type safety,
176low-level operations, flexibility, and the capability of representing
177'all' high-level languages cleanly. It is the common code
178representation used throughout all phases of the LLVM compilation
179strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000180</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000181
Chris Lattner00950542001-06-06 20:29:01 +0000182<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000183<div class="doc_section"> <a name="introduction">Introduction</a> </div>
184<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000185
Misha Brukman9d0919f2003-11-08 01:05:38 +0000186<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000187
Chris Lattner261efe92003-11-25 01:02:51 +0000188<p>The LLVM code representation is designed to be used in three
189different forms: as an in-memory compiler IR, as an on-disk bytecode
190representation (suitable for fast loading by a Just-In-Time compiler),
191and as a human readable assembly language representation. This allows
192LLVM to provide a powerful intermediate representation for efficient
193compiler transformations and analysis, while providing a natural means
194to debug and visualize the transformations. The three different forms
195of LLVM are all equivalent. This document describes the human readable
196representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000197
John Criswellc1f786c2005-05-13 22:25:59 +0000198<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000199while being expressive, typed, and extensible at the same time. It
200aims to be a "universal IR" of sorts, by being at a low enough level
201that high-level ideas may be cleanly mapped to it (similar to how
202microprocessors are "universal IR's", allowing many source languages to
203be mapped to them). By providing type information, LLVM can be used as
204the target of optimizations: for example, through pointer analysis, it
205can be proven that a C automatic variable is never accessed outside of
206the current function... allowing it to be promoted to a simple SSA
207value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000208
Misha Brukman9d0919f2003-11-08 01:05:38 +0000209</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000210
Chris Lattner00950542001-06-06 20:29:01 +0000211<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000212<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000213
Misha Brukman9d0919f2003-11-08 01:05:38 +0000214<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000215
Chris Lattner261efe92003-11-25 01:02:51 +0000216<p>It is important to note that this document describes 'well formed'
217LLVM assembly language. There is a difference between what the parser
218accepts and what is considered 'well formed'. For example, the
219following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000220
221<pre>
222 %x = <a href="#i_add">add</a> int 1, %x
223</pre>
224
Chris Lattner261efe92003-11-25 01:02:51 +0000225<p>...because the definition of <tt>%x</tt> does not dominate all of
226its uses. The LLVM infrastructure provides a verification pass that may
227be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000228automatically run by the parser after parsing input assembly and by
Chris Lattner261efe92003-11-25 01:02:51 +0000229the optimizer before it outputs bytecode. The violations pointed out
230by the verifier pass indicate bugs in transformation passes or input to
231the parser.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000232
Chris Lattner261efe92003-11-25 01:02:51 +0000233<!-- Describe the typesetting conventions here. --> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000234
Chris Lattner00950542001-06-06 20:29:01 +0000235<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000236<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000237<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000238
Misha Brukman9d0919f2003-11-08 01:05:38 +0000239<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000240
Chris Lattner261efe92003-11-25 01:02:51 +0000241<p>LLVM uses three different forms of identifiers, for different
242purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000243
Chris Lattner00950542001-06-06 20:29:01 +0000244<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000245 <li>Named values are represented as a string of characters with a '%' prefix.
246 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
247 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
248 Identifiers which require other characters in their names can be surrounded
249 with quotes. In this way, anything except a <tt>"</tt> character can be used
250 in a name.</li>
251
252 <li>Unnamed values are represented as an unsigned numeric value with a '%'
253 prefix. For example, %12, %2, %44.</li>
254
Reid Spencercc16dc32004-12-09 18:02:53 +0000255 <li>Constants, which are described in a <a href="#constants">section about
256 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000257</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000258
259<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
260don't need to worry about name clashes with reserved words, and the set of
261reserved words may be expanded in the future without penalty. Additionally,
262unnamed identifiers allow a compiler to quickly come up with a temporary
263variable without having to avoid symbol table conflicts.</p>
264
Chris Lattner261efe92003-11-25 01:02:51 +0000265<p>Reserved words in LLVM are very similar to reserved words in other
266languages. There are keywords for different opcodes ('<tt><a
Chris Lattnere5d947b2004-12-09 16:36:40 +0000267href="#i_add">add</a></tt>', '<tt><a href="#i_cast">cast</a></tt>', '<tt><a
268href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
269href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
270and others. These reserved words cannot conflict with variable names, because
271none of them start with a '%' character.</p>
272
273<p>Here is an example of LLVM code to multiply the integer variable
274'<tt>%X</tt>' by 8:</p>
275
Misha Brukman9d0919f2003-11-08 01:05:38 +0000276<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000277
278<pre>
279 %result = <a href="#i_mul">mul</a> uint %X, 8
280</pre>
281
Misha Brukman9d0919f2003-11-08 01:05:38 +0000282<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000283
284<pre>
285 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
286</pre>
287
Misha Brukman9d0919f2003-11-08 01:05:38 +0000288<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000289
290<pre>
291 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
292 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
293 %result = <a href="#i_add">add</a> uint %1, %1
294</pre>
295
Chris Lattner261efe92003-11-25 01:02:51 +0000296<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
297important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000298
Chris Lattner00950542001-06-06 20:29:01 +0000299<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000300
301 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
302 line.</li>
303
304 <li>Unnamed temporaries are created when the result of a computation is not
305 assigned to a named value.</li>
306
Misha Brukman9d0919f2003-11-08 01:05:38 +0000307 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000308
Misha Brukman9d0919f2003-11-08 01:05:38 +0000309</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000310
John Criswelle4c57cc2005-05-12 16:52:32 +0000311<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000312demonstrating instructions, we will follow an instruction with a comment that
313defines the type and name of value produced. Comments are shown in italic
314text.</p>
315
Misha Brukman9d0919f2003-11-08 01:05:38 +0000316</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000317
318<!-- *********************************************************************** -->
319<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
320<!-- *********************************************************************** -->
321
322<!-- ======================================================================= -->
323<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
324</div>
325
326<div class="doc_text">
327
328<p>LLVM programs are composed of "Module"s, each of which is a
329translation unit of the input programs. Each module consists of
330functions, global variables, and symbol table entries. Modules may be
331combined together with the LLVM linker, which merges function (and
332global variable) definitions, resolves forward declarations, and merges
333symbol table entries. Here is an example of the "hello world" module:</p>
334
335<pre><i>; Declare the string constant as a global constant...</i>
336<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
337 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
338
339<i>; External declaration of the puts function</i>
340<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
341
342<i>; Definition of main function</i>
343int %main() { <i>; int()* </i>
344 <i>; Convert [13x sbyte]* to sbyte *...</i>
345 %cast210 = <a
346 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
347
348 <i>; Call puts function to write out the string to stdout...</i>
349 <a
350 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
351 <a
352 href="#i_ret">ret</a> int 0<br>}<br></pre>
353
354<p>This example is made up of a <a href="#globalvars">global variable</a>
355named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
356function, and a <a href="#functionstructure">function definition</a>
357for "<tt>main</tt>".</p>
358
Chris Lattnere5d947b2004-12-09 16:36:40 +0000359<p>In general, a module is made up of a list of global values,
360where both functions and global variables are global values. Global values are
361represented by a pointer to a memory location (in this case, a pointer to an
362array of char, and a pointer to a function), and have one of the following <a
363href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000364
Chris Lattnere5d947b2004-12-09 16:36:40 +0000365</div>
366
367<!-- ======================================================================= -->
368<div class="doc_subsection">
369 <a name="linkage">Linkage Types</a>
370</div>
371
372<div class="doc_text">
373
374<p>
375All Global Variables and Functions have one of the following types of linkage:
376</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000377
378<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000379
Chris Lattnerfa730212004-12-09 16:11:40 +0000380 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000381
382 <dd>Global values with internal linkage are only directly accessible by
383 objects in the current module. In particular, linking code into a module with
384 an internal global value may cause the internal to be renamed as necessary to
385 avoid collisions. Because the symbol is internal to the module, all
386 references can be updated. This corresponds to the notion of the
387 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattnerfa730212004-12-09 16:11:40 +0000388 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000389
Chris Lattnerfa730212004-12-09 16:11:40 +0000390 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000391
392 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
393 the twist that linking together two modules defining the same
394 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
395 is typically used to implement inline functions. Unreferenced
396 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000397 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000398
Chris Lattnerfa730212004-12-09 16:11:40 +0000399 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000400
401 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
402 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
403 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000404 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000405
Chris Lattnerfa730212004-12-09 16:11:40 +0000406 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000407
408 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
409 pointer to array type. When two global variables with appending linkage are
410 linked together, the two global arrays are appended together. This is the
411 LLVM, typesafe, equivalent of having the system linker append together
412 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000413 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000414
Chris Lattnerfa730212004-12-09 16:11:40 +0000415 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000416
417 <dd>If none of the above identifiers are used, the global is externally
418 visible, meaning that it participates in linkage and can be used to resolve
419 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000420 </dd>
421</dl>
422
Chris Lattnerfa730212004-12-09 16:11:40 +0000423<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
424variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
425variable and was linked with this one, one of the two would be renamed,
426preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
427external (i.e., lacking any linkage declarations), they are accessible
428outside of the current module. It is illegal for a function <i>declaration</i>
429to have any linkage type other than "externally visible".</a></p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000430
Chris Lattnerfa730212004-12-09 16:11:40 +0000431</div>
432
433<!-- ======================================================================= -->
434<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000435 <a name="callingconv">Calling Conventions</a>
436</div>
437
438<div class="doc_text">
439
440<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
441and <a href="#i_invoke">invokes</a> can all have an optional calling convention
442specified for the call. The calling convention of any pair of dynamic
443caller/callee must match, or the behavior of the program is undefined. The
444following calling conventions are supported by LLVM, and more may be added in
445the future:</p>
446
447<dl>
448 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
449
450 <dd>This calling convention (the default if no other calling convention is
451 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000452 supports varargs function calls and tolerates some mismatch in the declared
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000453 prototype and implemented declaration of the function (as does normal C).
454 </dd>
455
456 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
457
458 <dd>This calling convention attempts to make calls as fast as possible
459 (e.g. by passing things in registers). This calling convention allows the
460 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000461 without having to conform to an externally specified ABI. Implementations of
462 this convention should allow arbitrary tail call optimization to be supported.
463 This calling convention does not support varargs and requires the prototype of
464 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000465 </dd>
466
467 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
468
469 <dd>This calling convention attempts to make code in the caller as efficient
470 as possible under the assumption that the call is not commonly executed. As
471 such, these calls often preserve all registers so that the call does not break
472 any live ranges in the caller side. This calling convention does not support
473 varargs and requires the prototype of all callees to exactly match the
474 prototype of the function definition.
475 </dd>
476
Chris Lattnercfe6b372005-05-07 01:46:40 +0000477 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000478
479 <dd>Any calling convention may be specified by number, allowing
480 target-specific calling conventions to be used. Target specific calling
481 conventions start at 64.
482 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000483</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000484
485<p>More calling conventions can be added/defined on an as-needed basis, to
486support pascal conventions or any other well-known target-independent
487convention.</p>
488
489</div>
490
491<!-- ======================================================================= -->
492<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000493 <a name="globalvars">Global Variables</a>
494</div>
495
496<div class="doc_text">
497
Chris Lattner3689a342005-02-12 19:30:21 +0000498<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000499instead of run-time. Global variables may optionally be initialized, may have
500an explicit section to be placed in, and may
Chris Lattner2cbdc452005-11-06 08:02:57 +0000501have an optional explicit alignment specified. A
John Criswell0ec250c2005-10-24 16:17:18 +0000502variable may be defined as a global "constant," which indicates that the
Chris Lattner3689a342005-02-12 19:30:21 +0000503contents of the variable will <b>never</b> be modified (enabling better
504optimization, allowing the global data to be placed in the read-only section of
505an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000506cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000507
508<p>
509LLVM explicitly allows <em>declarations</em> of global variables to be marked
510constant, even if the final definition of the global is not. This capability
511can be used to enable slightly better optimization of the program, but requires
512the language definition to guarantee that optimizations based on the
513'constantness' are valid for the translation units that do not include the
514definition.
515</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000516
517<p>As SSA values, global variables define pointer values that are in
518scope (i.e. they dominate) all basic blocks in the program. Global
519variables always define a pointer to their "content" type because they
520describe a region of memory, and all memory objects in LLVM are
521accessed through pointers.</p>
522
Chris Lattner88f6c462005-11-12 00:45:07 +0000523<p>LLVM allows an explicit section to be specified for globals. If the target
524supports it, it will emit globals to the section specified.</p>
525
Chris Lattner2cbdc452005-11-06 08:02:57 +0000526<p>An explicit alignment may be specified for a global. If not present, or if
527the alignment is set to zero, the alignment of the global is set by the target
528to whatever it feels convenient. If an explicit alignment is specified, the
529global is forced to have at least that much alignment. All alignments must be
530a power of 2.</p>
531
Chris Lattnerfa730212004-12-09 16:11:40 +0000532</div>
533
534
535<!-- ======================================================================= -->
536<div class="doc_subsection">
537 <a name="functionstructure">Functions</a>
538</div>
539
540<div class="doc_text">
541
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000542<p>LLVM function definitions consist of an optional <a href="#linkage">linkage
543type</a>, an optional <a href="#callingconv">calling convention</a>, a return
Chris Lattner88f6c462005-11-12 00:45:07 +0000544type, a function name, a (possibly empty) argument list, an optional section,
545an optional alignment, an opening curly brace,
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000546a list of basic blocks, and a closing curly brace. LLVM function declarations
547are defined with the "<tt>declare</tt>" keyword, an optional <a
Chris Lattner2cbdc452005-11-06 08:02:57 +0000548href="#callingconv">calling convention</a>, a return type, a function name,
549a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000550
551<p>A function definition contains a list of basic blocks, forming the CFG for
552the function. Each basic block may optionally start with a label (giving the
553basic block a symbol table entry), contains a list of instructions, and ends
554with a <a href="#terminators">terminator</a> instruction (such as a branch or
555function return).</p>
556
John Criswelle4c57cc2005-05-12 16:52:32 +0000557<p>The first basic block in a program is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000558executed on entrance to the function, and it is not allowed to have predecessor
559basic blocks (i.e. there can not be any branches to the entry block of a
560function). Because the block can have no predecessors, it also cannot have any
561<a href="#i_phi">PHI nodes</a>.</p>
562
563<p>LLVM functions are identified by their name and type signature. Hence, two
564functions with the same name but different parameter lists or return values are
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000565considered different functions, and LLVM will resolve references to each
Chris Lattnerfa730212004-12-09 16:11:40 +0000566appropriately.</p>
567
Chris Lattner88f6c462005-11-12 00:45:07 +0000568<p>LLVM allows an explicit section to be specified for functions. If the target
569supports it, it will emit functions to the section specified.</p>
570
Chris Lattner2cbdc452005-11-06 08:02:57 +0000571<p>An explicit alignment may be specified for a function. If not present, or if
572the alignment is set to zero, the alignment of the function is set by the target
573to whatever it feels convenient. If an explicit alignment is specified, the
574function is forced to have at least that much alignment. All alignments must be
575a power of 2.</p>
576
Chris Lattnerfa730212004-12-09 16:11:40 +0000577</div>
578
579
580
Chris Lattner00950542001-06-06 20:29:01 +0000581<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000582<div class="doc_section"> <a name="typesystem">Type System</a> </div>
583<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000584
Misha Brukman9d0919f2003-11-08 01:05:38 +0000585<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000586
Misha Brukman9d0919f2003-11-08 01:05:38 +0000587<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000588intermediate representation. Being typed enables a number of
589optimizations to be performed on the IR directly, without having to do
590extra analyses on the side before the transformation. A strong type
591system makes it easier to read the generated code and enables novel
592analyses and transformations that are not feasible to perform on normal
593three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000594
595</div>
596
Chris Lattner00950542001-06-06 20:29:01 +0000597<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000598<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000599<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000600<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000601system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000602
Reid Spencerd3f876c2004-11-01 08:19:36 +0000603<table class="layout">
604 <tr class="layout">
605 <td class="left">
606 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000607 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000608 <tr><th>Type</th><th>Description</th></tr>
609 <tr><td><tt>void</tt></td><td>No value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000610 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
611 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
612 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
613 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
614 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000615 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000616 </tbody>
617 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000618 </td>
619 <td class="right">
620 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000621 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000622 <tr><th>Type</th><th>Description</th></tr>
623 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000624 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
625 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
626 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
627 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
628 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000629 </tbody>
630 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000631 </td>
632 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000633</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000634</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000635
Chris Lattner00950542001-06-06 20:29:01 +0000636<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000637<div class="doc_subsubsection"> <a name="t_classifications">Type
638Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000639<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000640<p>These different primitive types fall into a few useful
641classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000642
643<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000644 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000645 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000646 <tr>
647 <td><a name="t_signed">signed</a></td>
648 <td><tt>sbyte, short, int, long, float, double</tt></td>
649 </tr>
650 <tr>
651 <td><a name="t_unsigned">unsigned</a></td>
652 <td><tt>ubyte, ushort, uint, ulong</tt></td>
653 </tr>
654 <tr>
655 <td><a name="t_integer">integer</a></td>
656 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
657 </tr>
658 <tr>
659 <td><a name="t_integral">integral</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000660 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
661 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000662 </tr>
663 <tr>
664 <td><a name="t_floating">floating point</a></td>
665 <td><tt>float, double</tt></td>
666 </tr>
667 <tr>
668 <td><a name="t_firstclass">first class</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000669 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
670 float, double, <a href="#t_pointer">pointer</a>,
671 <a href="#t_packed">packed</a></tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000672 </tr>
673 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000674</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000675
Chris Lattner261efe92003-11-25 01:02:51 +0000676<p>The <a href="#t_firstclass">first class</a> types are perhaps the
677most important. Values of these types are the only ones which can be
678produced by instructions, passed as arguments, or used as operands to
679instructions. This means that all structures and arrays must be
680manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000681</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000682
Chris Lattner00950542001-06-06 20:29:01 +0000683<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000684<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000685
Misha Brukman9d0919f2003-11-08 01:05:38 +0000686<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000687
Chris Lattner261efe92003-11-25 01:02:51 +0000688<p>The real power in LLVM comes from the derived types in the system.
689This is what allows a programmer to represent arrays, functions,
690pointers, and other useful types. Note that these derived types may be
691recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000692
Misha Brukman9d0919f2003-11-08 01:05:38 +0000693</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000694
Chris Lattner00950542001-06-06 20:29:01 +0000695<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000696<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000697
Misha Brukman9d0919f2003-11-08 01:05:38 +0000698<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000699
Chris Lattner00950542001-06-06 20:29:01 +0000700<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000701
Misha Brukman9d0919f2003-11-08 01:05:38 +0000702<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000703sequentially in memory. The array type requires a size (number of
704elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000705
Chris Lattner7faa8832002-04-14 06:13:44 +0000706<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000707
708<pre>
709 [&lt;# elements&gt; x &lt;elementtype&gt;]
710</pre>
711
John Criswelle4c57cc2005-05-12 16:52:32 +0000712<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000713be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000714
Chris Lattner7faa8832002-04-14 06:13:44 +0000715<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000716<table class="layout">
717 <tr class="layout">
718 <td class="left">
719 <tt>[40 x int ]</tt><br/>
720 <tt>[41 x int ]</tt><br/>
721 <tt>[40 x uint]</tt><br/>
722 </td>
723 <td class="left">
724 Array of 40 integer values.<br/>
725 Array of 41 integer values.<br/>
726 Array of 40 unsigned integer values.<br/>
727 </td>
728 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000729</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000730<p>Here are some examples of multidimensional arrays:</p>
731<table class="layout">
732 <tr class="layout">
733 <td class="left">
734 <tt>[3 x [4 x int]]</tt><br/>
735 <tt>[12 x [10 x float]]</tt><br/>
736 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
737 </td>
738 <td class="left">
John Criswellc1f786c2005-05-13 22:25:59 +0000739 3x4 array of integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000740 12x10 array of single precision floating point values.<br/>
741 2x3x4 array of unsigned integer values.<br/>
742 </td>
743 </tr>
744</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000745
John Criswell0ec250c2005-10-24 16:17:18 +0000746<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
747length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000748LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
749As a special case, however, zero length arrays are recognized to be variable
750length. This allows implementation of 'pascal style arrays' with the LLVM
751type "{ int, [0 x float]}", for example.</p>
752
Misha Brukman9d0919f2003-11-08 01:05:38 +0000753</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000754
Chris Lattner00950542001-06-06 20:29:01 +0000755<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000756<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000757<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000758<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000759<p>The function type can be thought of as a function signature. It
760consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000761Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000762(which are structures of pointers to functions), for indirect function
763calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000764<p>
765The return type of a function type cannot be an aggregate type.
766</p>
Chris Lattner00950542001-06-06 20:29:01 +0000767<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000768<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000769<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +0000770specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000771which indicates that the function takes a variable number of arguments.
772Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000773 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000774<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000775<table class="layout">
776 <tr class="layout">
777 <td class="left">
778 <tt>int (int)</tt> <br/>
779 <tt>float (int, int *) *</tt><br/>
780 <tt>int (sbyte *, ...)</tt><br/>
781 </td>
782 <td class="left">
783 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
784 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukmanc24b7582004-08-12 20:16:08 +0000785 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
Reid Spencerd3f876c2004-11-01 08:19:36 +0000786 returning <tt>float</tt>.<br/>
787 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
788 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
789 the signature for <tt>printf</tt> in LLVM.<br/>
790 </td>
791 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000792</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000793
Misha Brukman9d0919f2003-11-08 01:05:38 +0000794</div>
Chris Lattner00950542001-06-06 20:29:01 +0000795<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000796<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000797<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000798<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000799<p>The structure type is used to represent a collection of data members
800together in memory. The packing of the field types is defined to match
801the ABI of the underlying processor. The elements of a structure may
802be any type that has a size.</p>
803<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
804and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
805field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
806instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000807<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000808<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000809<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000810<table class="layout">
811 <tr class="layout">
812 <td class="left">
813 <tt>{ int, int, int }</tt><br/>
814 <tt>{ float, int (int) * }</tt><br/>
815 </td>
816 <td class="left">
817 a triple of three <tt>int</tt> values<br/>
818 A pair, where the first element is a <tt>float</tt> and the second element
819 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
820 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
821 </td>
822 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000823</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000824</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000825
Chris Lattner00950542001-06-06 20:29:01 +0000826<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000827<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000828<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000829<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000830<p>As in many languages, the pointer type represents a pointer or
831reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000832<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000833<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000834<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000835<table class="layout">
836 <tr class="layout">
837 <td class="left">
838 <tt>[4x int]*</tt><br/>
839 <tt>int (int *) *</tt><br/>
840 </td>
841 <td class="left">
842 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
843 four <tt>int</tt> values<br/>
844 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattnera977c482005-02-19 02:22:14 +0000845 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerd3f876c2004-11-01 08:19:36 +0000846 <tt>int</tt>.<br/>
847 </td>
848 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000849</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000850</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000851
Chris Lattnera58561b2004-08-12 19:12:28 +0000852<!-- _______________________________________________________________________ -->
853<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000854<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000855
Chris Lattnera58561b2004-08-12 19:12:28 +0000856<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000857
Chris Lattnera58561b2004-08-12 19:12:28 +0000858<p>A packed type is a simple derived type that represents a vector
859of elements. Packed types are used when multiple primitive data
860are operated in parallel using a single instruction (SIMD).
861A packed type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000862elements) and an underlying primitive data type. Vectors must have a power
863of two length (1, 2, 4, 8, 16 ...). Packed types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000864considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000865
Chris Lattnera58561b2004-08-12 19:12:28 +0000866<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000867
868<pre>
869 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
870</pre>
871
John Criswellc1f786c2005-05-13 22:25:59 +0000872<p>The number of elements is a constant integer value; elementtype may
Chris Lattnera58561b2004-08-12 19:12:28 +0000873be any integral or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000874
Chris Lattnera58561b2004-08-12 19:12:28 +0000875<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000876
Reid Spencerd3f876c2004-11-01 08:19:36 +0000877<table class="layout">
878 <tr class="layout">
879 <td class="left">
880 <tt>&lt;4 x int&gt;</tt><br/>
881 <tt>&lt;8 x float&gt;</tt><br/>
882 <tt>&lt;2 x uint&gt;</tt><br/>
883 </td>
884 <td class="left">
885 Packed vector of 4 integer values.<br/>
886 Packed vector of 8 floating-point values.<br/>
887 Packed vector of 2 unsigned integer values.<br/>
888 </td>
889 </tr>
890</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000891</div>
892
Chris Lattner69c11bb2005-04-25 17:34:15 +0000893<!-- _______________________________________________________________________ -->
894<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
895<div class="doc_text">
896
897<h5>Overview:</h5>
898
899<p>Opaque types are used to represent unknown types in the system. This
900corresponds (for example) to the C notion of a foward declared structure type.
901In LLVM, opaque types can eventually be resolved to any type (not just a
902structure type).</p>
903
904<h5>Syntax:</h5>
905
906<pre>
907 opaque
908</pre>
909
910<h5>Examples:</h5>
911
912<table class="layout">
913 <tr class="layout">
914 <td class="left">
915 <tt>opaque</tt>
916 </td>
917 <td class="left">
918 An opaque type.<br/>
919 </td>
920 </tr>
921</table>
922</div>
923
924
Chris Lattnerc3f59762004-12-09 17:30:23 +0000925<!-- *********************************************************************** -->
926<div class="doc_section"> <a name="constants">Constants</a> </div>
927<!-- *********************************************************************** -->
928
929<div class="doc_text">
930
931<p>LLVM has several different basic types of constants. This section describes
932them all and their syntax.</p>
933
934</div>
935
936<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000937<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000938
939<div class="doc_text">
940
941<dl>
942 <dt><b>Boolean constants</b></dt>
943
944 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
945 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
946 </dd>
947
948 <dt><b>Integer constants</b></dt>
949
Reid Spencercc16dc32004-12-09 18:02:53 +0000950 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000951 href="#t_integer">integer</a> type. Negative numbers may be used with signed
952 integer types.
953 </dd>
954
955 <dt><b>Floating point constants</b></dt>
956
957 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
958 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000959 notation (see below). Floating point constants must have a <a
960 href="#t_floating">floating point</a> type. </dd>
961
962 <dt><b>Null pointer constants</b></dt>
963
John Criswell9e2485c2004-12-10 15:51:16 +0000964 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +0000965 and must be of <a href="#t_pointer">pointer type</a>.</dd>
966
967</dl>
968
John Criswell9e2485c2004-12-10 15:51:16 +0000969<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +0000970of floating point constants. For example, the form '<tt>double
9710x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
9724.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +0000973(and the only time that they are generated by the disassembler) is when a
974floating point constant must be emitted but it cannot be represented as a
975decimal floating point number. For example, NaN's, infinities, and other
976special values are represented in their IEEE hexadecimal format so that
977assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000978
979</div>
980
981<!-- ======================================================================= -->
982<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
983</div>
984
985<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000986<p>Aggregate constants arise from aggregation of simple constants
987and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000988
989<dl>
990 <dt><b>Structure constants</b></dt>
991
992 <dd>Structure constants are represented with notation similar to structure
993 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000994 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
995 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
996 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +0000997 types of elements must match those specified by the type.
998 </dd>
999
1000 <dt><b>Array constants</b></dt>
1001
1002 <dd>Array constants are represented with notation similar to array type
1003 definitions (a comma separated list of elements, surrounded by square brackets
John Criswell9e2485c2004-12-10 15:51:16 +00001004 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +00001005 constants must have <a href="#t_array">array type</a>, and the number and
1006 types of elements must match those specified by the type.
1007 </dd>
1008
1009 <dt><b>Packed constants</b></dt>
1010
1011 <dd>Packed constants are represented with notation similar to packed type
1012 definitions (a comma separated list of elements, surrounded by
John Criswell9e2485c2004-12-10 15:51:16 +00001013 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; int 42,
Chris Lattnerc3f59762004-12-09 17:30:23 +00001014 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
1015 href="#t_packed">packed type</a>, and the number and types of elements must
1016 match those specified by the type.
1017 </dd>
1018
1019 <dt><b>Zero initialization</b></dt>
1020
1021 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1022 value to zero of <em>any</em> type, including scalar and aggregate types.
1023 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +00001024 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +00001025 initializers.
1026 </dd>
1027</dl>
1028
1029</div>
1030
1031<!-- ======================================================================= -->
1032<div class="doc_subsection">
1033 <a name="globalconstants">Global Variable and Function Addresses</a>
1034</div>
1035
1036<div class="doc_text">
1037
1038<p>The addresses of <a href="#globalvars">global variables</a> and <a
1039href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +00001040constants. These constants are explicitly referenced when the <a
1041href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +00001042href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1043file:</p>
1044
1045<pre>
1046 %X = global int 17
1047 %Y = global int 42
1048 %Z = global [2 x int*] [ int* %X, int* %Y ]
1049</pre>
1050
1051</div>
1052
1053<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +00001054<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001055<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +00001056 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +00001057 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +00001058 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001059
Reid Spencer2dc45b82004-12-09 18:13:12 +00001060 <p>Undefined values indicate to the compiler that the program is well defined
1061 no matter what value is used, giving the compiler more freedom to optimize.
1062 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001063</div>
1064
1065<!-- ======================================================================= -->
1066<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1067</div>
1068
1069<div class="doc_text">
1070
1071<p>Constant expressions are used to allow expressions involving other constants
1072to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +00001073href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +00001074that does not have side effects (e.g. load and call are not supported). The
1075following is the syntax for constant expressions:</p>
1076
1077<dl>
1078 <dt><b><tt>cast ( CST to TYPE )</tt></b></dt>
1079
1080 <dd>Cast a constant to another type.</dd>
1081
1082 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1083
1084 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1085 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1086 instruction, the index list may have zero or more indexes, which are required
1087 to make sense for the type of "CSTPTR".</dd>
1088
1089 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1090
Reid Spencer2dc45b82004-12-09 18:13:12 +00001091 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1092 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +00001093 binary</a> operations. The constraints on operands are the same as those for
1094 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +00001095 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001096</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001097</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00001098
Chris Lattner00950542001-06-06 20:29:01 +00001099<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00001100<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1101<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00001102
Misha Brukman9d0919f2003-11-08 01:05:38 +00001103<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001104
Chris Lattner261efe92003-11-25 01:02:51 +00001105<p>The LLVM instruction set consists of several different
1106classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +00001107instructions</a>, <a href="#binaryops">binary instructions</a>,
1108<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +00001109 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1110instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001111
Misha Brukman9d0919f2003-11-08 01:05:38 +00001112</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001113
Chris Lattner00950542001-06-06 20:29:01 +00001114<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001115<div class="doc_subsection"> <a name="terminators">Terminator
1116Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001117
Misha Brukman9d0919f2003-11-08 01:05:38 +00001118<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001119
Chris Lattner261efe92003-11-25 01:02:51 +00001120<p>As mentioned <a href="#functionstructure">previously</a>, every
1121basic block in a program ends with a "Terminator" instruction, which
1122indicates which block should be executed after the current block is
1123finished. These terminator instructions typically yield a '<tt>void</tt>'
1124value: they produce control flow, not values (the one exception being
1125the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +00001126<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +00001127 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1128instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +00001129the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1130 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1131 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001132
Misha Brukman9d0919f2003-11-08 01:05:38 +00001133</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001134
Chris Lattner00950542001-06-06 20:29:01 +00001135<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001136<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1137Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001138<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001139<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001140<pre> ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001141 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001142</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001143<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001144<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +00001145value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +00001146<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +00001147returns a value and then causes control flow, and one that just causes
1148control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001149<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001150<p>The '<tt>ret</tt>' instruction may return any '<a
1151 href="#t_firstclass">first class</a>' type. Notice that a function is
1152not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1153instruction inside of the function that returns a value that does not
1154match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001155<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001156<p>When the '<tt>ret</tt>' instruction is executed, control flow
1157returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +00001158 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +00001159the instruction after the call. If the caller was an "<a
1160 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +00001161at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +00001162returns a value, that value shall set the call or invoke instruction's
1163return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001164<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001165<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001166 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001167</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001168</div>
Chris Lattner00950542001-06-06 20:29:01 +00001169<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001170<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001171<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001172<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001173<pre> br bool &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner00950542001-06-06 20:29:01 +00001174</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001175<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001176<p>The '<tt>br</tt>' instruction is used to cause control flow to
1177transfer to a different basic block in the current function. There are
1178two forms of this instruction, corresponding to a conditional branch
1179and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001180<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001181<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1182single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1183unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1184value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001185<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001186<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1187argument is evaluated. If the value is <tt>true</tt>, control flows
1188to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1189control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001190<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001191<pre>Test:<br> %cond = <a href="#i_setcc">seteq</a> int %a, %b<br> br bool %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
1192 href="#i_ret">ret</a> int 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> int 0<br></pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001193</div>
Chris Lattner00950542001-06-06 20:29:01 +00001194<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001195<div class="doc_subsubsection">
1196 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1197</div>
1198
Misha Brukman9d0919f2003-11-08 01:05:38 +00001199<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001200<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001201
1202<pre>
1203 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1204</pre>
1205
Chris Lattner00950542001-06-06 20:29:01 +00001206<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001207
1208<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1209several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +00001210instruction, allowing a branch to occur to one of many possible
1211destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001212
1213
Chris Lattner00950542001-06-06 20:29:01 +00001214<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001215
1216<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1217comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1218an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1219table is not allowed to contain duplicate constant entries.</p>
1220
Chris Lattner00950542001-06-06 20:29:01 +00001221<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001222
Chris Lattner261efe92003-11-25 01:02:51 +00001223<p>The <tt>switch</tt> instruction specifies a table of values and
1224destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +00001225table is searched for the given value. If the value is found, control flow is
1226transfered to the corresponding destination; otherwise, control flow is
1227transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001228
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001229<h5>Implementation:</h5>
1230
1231<p>Depending on properties of the target machine and the particular
1232<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +00001233ways. For example, it could be generated as a series of chained conditional
1234branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001235
1236<h5>Example:</h5>
1237
1238<pre>
1239 <i>; Emulate a conditional br instruction</i>
1240 %Val = <a href="#i_cast">cast</a> bool %value to int
1241 switch int %Val, label %truedest [int 0, label %falsedest ]
1242
1243 <i>; Emulate an unconditional br instruction</i>
1244 switch uint 0, label %dest [ ]
1245
1246 <i>; Implement a jump table:</i>
1247 switch uint %val, label %otherwise [ uint 0, label %onzero
1248 uint 1, label %onone
1249 uint 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00001250</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001251</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001252
Chris Lattner00950542001-06-06 20:29:01 +00001253<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001254<div class="doc_subsubsection">
1255 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1256</div>
1257
Misha Brukman9d0919f2003-11-08 01:05:38 +00001258<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001259
Chris Lattner00950542001-06-06 20:29:01 +00001260<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001261
1262<pre>
1263 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)
1264 to label &lt;normal label&gt; except label &lt;exception label&gt;
1265</pre>
1266
Chris Lattner6536cfe2002-05-06 22:08:29 +00001267<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001268
1269<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1270function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +00001271'<tt>normal</tt>' label or the
1272'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001273"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1274"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +00001275href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1276continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001277
Chris Lattner00950542001-06-06 20:29:01 +00001278<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001279
Misha Brukman9d0919f2003-11-08 01:05:38 +00001280<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001281
Chris Lattner00950542001-06-06 20:29:01 +00001282<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001283 <li>
John Criswellc1f786c2005-05-13 22:25:59 +00001284 The optional "cconv" marker indicates which <a href="callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001285 convention</a> the call should use. If none is specified, the call defaults
1286 to using C calling conventions.
1287 </li>
1288 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1289 function value being invoked. In most cases, this is a direct function
1290 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1291 an arbitrary pointer to function value.
1292 </li>
1293
1294 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1295 function to be invoked. </li>
1296
1297 <li>'<tt>function args</tt>': argument list whose types match the function
1298 signature argument types. If the function signature indicates the function
1299 accepts a variable number of arguments, the extra arguments can be
1300 specified. </li>
1301
1302 <li>'<tt>normal label</tt>': the label reached when the called function
1303 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1304
1305 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1306 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1307
Chris Lattner00950542001-06-06 20:29:01 +00001308</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001309
Chris Lattner00950542001-06-06 20:29:01 +00001310<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001311
Misha Brukman9d0919f2003-11-08 01:05:38 +00001312<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001313href="#i_call">call</a></tt>' instruction in most regards. The primary
1314difference is that it establishes an association with a label, which is used by
1315the runtime library to unwind the stack.</p>
1316
1317<p>This instruction is used in languages with destructors to ensure that proper
1318cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1319exception. Additionally, this is important for implementation of
1320'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1321
Chris Lattner00950542001-06-06 20:29:01 +00001322<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001323<pre>
1324 %retval = invoke int %Test(int 15) to label %Continue
1325 except label %TestCleanup <i>; {int}:retval set</i>
1326 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
1327 except label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00001328</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001329</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001330
1331
Chris Lattner27f71f22003-09-03 00:41:47 +00001332<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00001333
Chris Lattner261efe92003-11-25 01:02:51 +00001334<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1335Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +00001336
Misha Brukman9d0919f2003-11-08 01:05:38 +00001337<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +00001338
Chris Lattner27f71f22003-09-03 00:41:47 +00001339<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001340<pre>
1341 unwind
1342</pre>
1343
Chris Lattner27f71f22003-09-03 00:41:47 +00001344<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001345
1346<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1347at the first callee in the dynamic call stack which used an <a
1348href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1349primarily used to implement exception handling.</p>
1350
Chris Lattner27f71f22003-09-03 00:41:47 +00001351<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001352
1353<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1354immediately halt. The dynamic call stack is then searched for the first <a
1355href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1356execution continues at the "exceptional" destination block specified by the
1357<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1358dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001359</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001360
1361<!-- _______________________________________________________________________ -->
1362
1363<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1364Instruction</a> </div>
1365
1366<div class="doc_text">
1367
1368<h5>Syntax:</h5>
1369<pre>
1370 unreachable
1371</pre>
1372
1373<h5>Overview:</h5>
1374
1375<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1376instruction is used to inform the optimizer that a particular portion of the
1377code is not reachable. This can be used to indicate that the code after a
1378no-return function cannot be reached, and other facts.</p>
1379
1380<h5>Semantics:</h5>
1381
1382<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1383</div>
1384
1385
1386
Chris Lattner00950542001-06-06 20:29:01 +00001387<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001388<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001389<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001390<p>Binary operators are used to do most of the computation in a
1391program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +00001392produce a single value. The operands might represent
Chris Lattnera58561b2004-08-12 19:12:28 +00001393multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1394The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +00001395necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001396<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001397</div>
Chris Lattner00950542001-06-06 20:29:01 +00001398<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001399<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1400Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001401<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001402<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001403<pre> &lt;result&gt; = add &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001404</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001405<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001406<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001407<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001408<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +00001409 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1410 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1411Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001412<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001413<p>The value produced is the integer or floating point sum of the two
1414operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001415<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001416<pre> &lt;result&gt; = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001417</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001418</div>
Chris Lattner00950542001-06-06 20:29:01 +00001419<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001420<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1421Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001422<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001423<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001424<pre> &lt;result&gt; = sub &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001425</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001426<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001427<p>The '<tt>sub</tt>' instruction returns the difference of its two
1428operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001429<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1430instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001431<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001432<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001433 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001434values.
1435This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1436Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001437<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001438<p>The value produced is the integer or floating point difference of
1439the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001440<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001441<pre> &lt;result&gt; = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001442 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1443</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001444</div>
Chris Lattner00950542001-06-06 20:29:01 +00001445<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001446<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1447Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001448<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001449<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001450<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001451</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001452<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001453<p>The '<tt>mul</tt>' instruction returns the product of its two
1454operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001455<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001456<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001457 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001458values.
1459This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1460Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001461<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001462<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +00001463two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001464<p>There is no signed vs unsigned multiplication. The appropriate
1465action is taken based on the type of the operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001466<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001467<pre> &lt;result&gt; = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001468</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001469</div>
Chris Lattner00950542001-06-06 20:29:01 +00001470<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001471<div class="doc_subsubsection"> <a name="i_div">'<tt>div</tt>'
1472Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001473<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001474<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001475<pre> &lt;result&gt; = div &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1476</pre>
1477<h5>Overview:</h5>
1478<p>The '<tt>div</tt>' instruction returns the quotient of its two
1479operands.</p>
1480<h5>Arguments:</h5>
1481<p>The two arguments to the '<tt>div</tt>' instruction must be either <a
1482 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001483values.
1484This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1485Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001486<h5>Semantics:</h5>
1487<p>The value produced is the integer or floating point quotient of the
1488two operands.</p>
1489<h5>Example:</h5>
1490<pre> &lt;result&gt; = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1491</pre>
1492</div>
1493<!-- _______________________________________________________________________ -->
1494<div class="doc_subsubsection"> <a name="i_rem">'<tt>rem</tt>'
1495Instruction</a> </div>
1496<div class="doc_text">
1497<h5>Syntax:</h5>
1498<pre> &lt;result&gt; = rem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1499</pre>
1500<h5>Overview:</h5>
1501<p>The '<tt>rem</tt>' instruction returns the remainder from the
1502division of its two operands.</p>
1503<h5>Arguments:</h5>
1504<p>The two arguments to the '<tt>rem</tt>' instruction must be either <a
1505 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001506values.
1507This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1508Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001509<h5>Semantics:</h5>
1510<p>This returns the <i>remainder</i> of a division (where the result
1511has the same sign as the divisor), not the <i>modulus</i> (where the
1512result has the same sign as the dividend) of a value. For more
John Criswell0ec250c2005-10-24 16:17:18 +00001513information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +00001514 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1515Math Forum</a>.</p>
1516<h5>Example:</h5>
1517<pre> &lt;result&gt; = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1518</pre>
1519</div>
1520<!-- _______________________________________________________________________ -->
1521<div class="doc_subsubsection"> <a name="i_setcc">'<tt>set<i>cc</i></tt>'
1522Instructions</a> </div>
1523<div class="doc_text">
1524<h5>Syntax:</h5>
1525<pre> &lt;result&gt; = seteq &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001526 &lt;result&gt; = setne &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1527 &lt;result&gt; = setlt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1528 &lt;result&gt; = setgt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1529 &lt;result&gt; = setle &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1530 &lt;result&gt; = setge &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1531</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001532<h5>Overview:</h5>
1533<p>The '<tt>set<i>cc</i></tt>' family of instructions returns a boolean
1534value based on a comparison of their two operands.</p>
1535<h5>Arguments:</h5>
1536<p>The two arguments to the '<tt>set<i>cc</i></tt>' instructions must
1537be of <a href="#t_firstclass">first class</a> type (it is not possible
1538to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1539or '<tt>void</tt>' values, etc...). Both arguments must have identical
1540types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001541<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001542<p>The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1543value if both operands are equal.<br>
1544The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1545value if both operands are unequal.<br>
1546The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1547value if the first operand is less than the second operand.<br>
1548The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1549value if the first operand is greater than the second operand.<br>
1550The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1551value if the first operand is less than or equal to the second operand.<br>
1552The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1553value if the first operand is greater than or equal to the second
1554operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001555<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001556<pre> &lt;result&gt; = seteq int 4, 5 <i>; yields {bool}:result = false</i>
Chris Lattner00950542001-06-06 20:29:01 +00001557 &lt;result&gt; = setne float 4, 5 <i>; yields {bool}:result = true</i>
1558 &lt;result&gt; = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1559 &lt;result&gt; = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1560 &lt;result&gt; = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1561 &lt;result&gt; = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1562</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001563</div>
Chris Lattner00950542001-06-06 20:29:01 +00001564<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001565<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1566Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001567<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001568<p>Bitwise binary operators are used to do various forms of
1569bit-twiddling in a program. They are generally very efficient
John Criswell9e2485c2004-12-10 15:51:16 +00001570instructions and can commonly be strength reduced from other
Chris Lattner261efe92003-11-25 01:02:51 +00001571instructions. They require two operands, execute an operation on them,
1572and produce a single value. The resulting value of the bitwise binary
1573operators is always the same type as its first operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001574</div>
Chris Lattner00950542001-06-06 20:29:01 +00001575<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001576<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1577Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001578<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001579<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001580<pre> &lt;result&gt; = and &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001581</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001582<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001583<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1584its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001585<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001586<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001587 href="#t_integral">integral</a> values. Both arguments must have
1588identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001589<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001590<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001591<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001592<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +00001593<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001594 <tbody>
1595 <tr>
1596 <td>In0</td>
1597 <td>In1</td>
1598 <td>Out</td>
1599 </tr>
1600 <tr>
1601 <td>0</td>
1602 <td>0</td>
1603 <td>0</td>
1604 </tr>
1605 <tr>
1606 <td>0</td>
1607 <td>1</td>
1608 <td>0</td>
1609 </tr>
1610 <tr>
1611 <td>1</td>
1612 <td>0</td>
1613 <td>0</td>
1614 </tr>
1615 <tr>
1616 <td>1</td>
1617 <td>1</td>
1618 <td>1</td>
1619 </tr>
1620 </tbody>
1621</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001622</div>
Chris Lattner00950542001-06-06 20:29:01 +00001623<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001624<pre> &lt;result&gt; = and int 4, %var <i>; yields {int}:result = 4 &amp; %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001625 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1626 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1627</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001628</div>
Chris Lattner00950542001-06-06 20:29:01 +00001629<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001630<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001631<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001632<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001633<pre> &lt;result&gt; = or &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001634</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001635<h5>Overview:</h5>
1636<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1637or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001638<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001639<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001640 href="#t_integral">integral</a> values. Both arguments must have
1641identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001642<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001643<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001644<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001645<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001646<table border="1" cellspacing="0" cellpadding="4">
1647 <tbody>
1648 <tr>
1649 <td>In0</td>
1650 <td>In1</td>
1651 <td>Out</td>
1652 </tr>
1653 <tr>
1654 <td>0</td>
1655 <td>0</td>
1656 <td>0</td>
1657 </tr>
1658 <tr>
1659 <td>0</td>
1660 <td>1</td>
1661 <td>1</td>
1662 </tr>
1663 <tr>
1664 <td>1</td>
1665 <td>0</td>
1666 <td>1</td>
1667 </tr>
1668 <tr>
1669 <td>1</td>
1670 <td>1</td>
1671 <td>1</td>
1672 </tr>
1673 </tbody>
1674</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001675</div>
Chris Lattner00950542001-06-06 20:29:01 +00001676<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001677<pre> &lt;result&gt; = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001678 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1679 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1680</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001681</div>
Chris Lattner00950542001-06-06 20:29:01 +00001682<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001683<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1684Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001685<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001686<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001687<pre> &lt;result&gt; = xor &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001688</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001689<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001690<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1691or of its two operands. The <tt>xor</tt> is used to implement the
1692"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001693<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001694<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001695 href="#t_integral">integral</a> values. Both arguments must have
1696identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001697<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001698<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001699<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001700<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001701<table border="1" cellspacing="0" cellpadding="4">
1702 <tbody>
1703 <tr>
1704 <td>In0</td>
1705 <td>In1</td>
1706 <td>Out</td>
1707 </tr>
1708 <tr>
1709 <td>0</td>
1710 <td>0</td>
1711 <td>0</td>
1712 </tr>
1713 <tr>
1714 <td>0</td>
1715 <td>1</td>
1716 <td>1</td>
1717 </tr>
1718 <tr>
1719 <td>1</td>
1720 <td>0</td>
1721 <td>1</td>
1722 </tr>
1723 <tr>
1724 <td>1</td>
1725 <td>1</td>
1726 <td>0</td>
1727 </tr>
1728 </tbody>
1729</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001730</div>
Chris Lattner261efe92003-11-25 01:02:51 +00001731<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +00001732<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001733<pre> &lt;result&gt; = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001734 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
1735 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner27f71f22003-09-03 00:41:47 +00001736 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00001737</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001738</div>
Chris Lattner00950542001-06-06 20:29:01 +00001739<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001740<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
1741Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001742<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001743<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001744<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001745</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001746<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001747<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
1748the left a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001749<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001750<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001751 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1752type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001753<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001754<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001755<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001756<pre> &lt;result&gt; = shl int 4, ubyte %var <i>; yields {int}:result = 4 &lt;&lt; %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001757 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1758 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1759</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001760</div>
Chris Lattner00950542001-06-06 20:29:01 +00001761<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001762<div class="doc_subsubsection"> <a name="i_shr">'<tt>shr</tt>'
1763Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001764<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001765<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001766<pre> &lt;result&gt; = shr &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001767</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001768<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001769<p>The '<tt>shr</tt>' instruction returns the first operand shifted to
1770the right a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001771<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001772<p>The first argument to the '<tt>shr</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001773 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1774type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001775<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001776<p>If the first argument is a <a href="#t_signed">signed</a> type, the
1777most significant bit is duplicated in the newly free'd bit positions.
1778If the first argument is unsigned, zero bits shall fill the empty
1779positions.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001780<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001781<pre> &lt;result&gt; = shr int 4, ubyte %var <i>; yields {int}:result = 4 &gt;&gt; %var</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +00001782 &lt;result&gt; = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
Chris Lattner00950542001-06-06 20:29:01 +00001783 &lt;result&gt; = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +00001784 &lt;result&gt; = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1785 &lt;result&gt; = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner00950542001-06-06 20:29:01 +00001786</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001787</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001788
Chris Lattner00950542001-06-06 20:29:01 +00001789<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001790<div class="doc_subsection">
1791 <a name="memoryops">Memory Access Operations</a>
1792</div>
1793
Misha Brukman9d0919f2003-11-08 01:05:38 +00001794<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001795
Chris Lattner261efe92003-11-25 01:02:51 +00001796<p>A key design point of an SSA-based representation is how it
1797represents memory. In LLVM, no memory locations are in SSA form, which
1798makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +00001799allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001800
Misha Brukman9d0919f2003-11-08 01:05:38 +00001801</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001802
Chris Lattner00950542001-06-06 20:29:01 +00001803<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001804<div class="doc_subsubsection">
1805 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
1806</div>
1807
Misha Brukman9d0919f2003-11-08 01:05:38 +00001808<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001809
Chris Lattner00950542001-06-06 20:29:01 +00001810<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001811
1812<pre>
1813 &lt;result&gt; = malloc &lt;type&gt;[, uint &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001814</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001815
Chris Lattner00950542001-06-06 20:29:01 +00001816<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001817
Chris Lattner261efe92003-11-25 01:02:51 +00001818<p>The '<tt>malloc</tt>' instruction allocates memory from the system
1819heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001820
Chris Lattner00950542001-06-06 20:29:01 +00001821<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001822
1823<p>The '<tt>malloc</tt>' instruction allocates
1824<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +00001825bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00001826appropriate type to the program. If "NumElements" is specified, it is the
1827number of elements allocated. If an alignment is specified, the value result
1828of the allocation is guaranteed to be aligned to at least that boundary. If
1829not specified, or if zero, the target can choose to align the allocation on any
1830convenient boundary.</p>
1831
Misha Brukman9d0919f2003-11-08 01:05:38 +00001832<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001833
Chris Lattner00950542001-06-06 20:29:01 +00001834<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001835
Chris Lattner261efe92003-11-25 01:02:51 +00001836<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
1837a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001838
Chris Lattner2cbdc452005-11-06 08:02:57 +00001839<h5>Example:</h5>
1840
1841<pre>
1842 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
1843
1844 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001845 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1846 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001847 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
1848 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +00001849</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001850</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001851
Chris Lattner00950542001-06-06 20:29:01 +00001852<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001853<div class="doc_subsubsection">
1854 <a name="i_free">'<tt>free</tt>' Instruction</a>
1855</div>
1856
Misha Brukman9d0919f2003-11-08 01:05:38 +00001857<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001858
Chris Lattner00950542001-06-06 20:29:01 +00001859<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001860
1861<pre>
1862 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +00001863</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001864
Chris Lattner00950542001-06-06 20:29:01 +00001865<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001866
Chris Lattner261efe92003-11-25 01:02:51 +00001867<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +00001868memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001869
Chris Lattner00950542001-06-06 20:29:01 +00001870<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001871
Chris Lattner261efe92003-11-25 01:02:51 +00001872<p>'<tt>value</tt>' shall be a pointer value that points to a value
1873that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
1874instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001875
Chris Lattner00950542001-06-06 20:29:01 +00001876<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001877
John Criswell9e2485c2004-12-10 15:51:16 +00001878<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +00001879after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001880
Chris Lattner00950542001-06-06 20:29:01 +00001881<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001882
1883<pre>
1884 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
Chris Lattner00950542001-06-06 20:29:01 +00001885 free [4 x ubyte]* %array
1886</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001887</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001888
Chris Lattner00950542001-06-06 20:29:01 +00001889<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001890<div class="doc_subsubsection">
1891 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
1892</div>
1893
Misha Brukman9d0919f2003-11-08 01:05:38 +00001894<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001895
Chris Lattner00950542001-06-06 20:29:01 +00001896<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001897
1898<pre>
1899 &lt;result&gt; = alloca &lt;type&gt;[, uint &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001900</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001901
Chris Lattner00950542001-06-06 20:29:01 +00001902<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001903
Chris Lattner261efe92003-11-25 01:02:51 +00001904<p>The '<tt>alloca</tt>' instruction allocates memory on the current
1905stack frame of the procedure that is live until the current function
1906returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001907
Chris Lattner00950542001-06-06 20:29:01 +00001908<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001909
John Criswell9e2485c2004-12-10 15:51:16 +00001910<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00001911bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00001912appropriate type to the program. If "NumElements" is specified, it is the
1913number of elements allocated. If an alignment is specified, the value result
1914of the allocation is guaranteed to be aligned to at least that boundary. If
1915not specified, or if zero, the target can choose to align the allocation on any
1916convenient boundary.</p>
1917
Misha Brukman9d0919f2003-11-08 01:05:38 +00001918<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001919
Chris Lattner00950542001-06-06 20:29:01 +00001920<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001921
John Criswellc1f786c2005-05-13 22:25:59 +00001922<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +00001923memory is automatically released when the function returns. The '<tt>alloca</tt>'
1924instruction is commonly used to represent automatic variables that must
1925have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +00001926 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001927instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001928
Chris Lattner00950542001-06-06 20:29:01 +00001929<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001930
1931<pre>
1932 %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001933 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001934 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
1935 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00001936</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001937</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001938
Chris Lattner00950542001-06-06 20:29:01 +00001939<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001940<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
1941Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001942<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00001943<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001944<pre> &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;<br> &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;<br></pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001945<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001946<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001947<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001948<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +00001949address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +00001950 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +00001951marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +00001952the number or order of execution of this <tt>load</tt> with other
1953volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
1954instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001955<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001956<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001957<h5>Examples:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001958<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1959 <a
1960 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001961 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1962</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001963</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001964<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001965<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
1966Instruction</a> </div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001967<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001968<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt; <i>; yields {void}</i>
Chris Lattnerf0651072003-09-08 18:27:49 +00001969 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt; <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001970</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001971<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001972<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001973<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001974<p>There are two arguments to the '<tt>store</tt>' instruction: a value
John Criswell0ec250c2005-10-24 16:17:18 +00001975to store and an address in which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00001976operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +00001977operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +00001978optimizer is not allowed to modify the number or order of execution of
1979this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
1980 href="#i_store">store</a></tt> instructions.</p>
1981<h5>Semantics:</h5>
1982<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
1983at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001984<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001985<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1986 <a
1987 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001988 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1989</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001990<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +00001991<div class="doc_subsubsection">
1992 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
1993</div>
1994
Misha Brukman9d0919f2003-11-08 01:05:38 +00001995<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +00001996<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00001997<pre>
1998 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
1999</pre>
2000
Chris Lattner7faa8832002-04-14 06:13:44 +00002001<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002002
2003<p>
2004The '<tt>getelementptr</tt>' instruction is used to get the address of a
2005subelement of an aggregate data structure.</p>
2006
Chris Lattner7faa8832002-04-14 06:13:44 +00002007<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002008
2009<p>This instruction takes a list of integer constants that indicate what
2010elements of the aggregate object to index to. The actual types of the arguments
2011provided depend on the type of the first pointer argument. The
2012'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +00002013levels of a structure or to a specific index in an array. When indexing into a
2014structure, only <tt>uint</tt>
John Criswellc1f786c2005-05-13 22:25:59 +00002015integer constants are allowed. When indexing into an array or pointer,
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002016<tt>int</tt> and <tt>long</tt> indexes are allowed of any sign.</p>
2017
Chris Lattner261efe92003-11-25 01:02:51 +00002018<p>For example, let's consider a C code fragment and how it gets
2019compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002020
2021<pre>
2022 struct RT {
2023 char A;
2024 int B[10][20];
2025 char C;
2026 };
2027 struct ST {
2028 int X;
2029 double Y;
2030 struct RT Z;
2031 };
2032
2033 int *foo(struct ST *s) {
2034 return &amp;s[1].Z.B[5][13];
2035 }
2036</pre>
2037
Misha Brukman9d0919f2003-11-08 01:05:38 +00002038<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002039
2040<pre>
2041 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2042 %ST = type { int, double, %RT }
2043
Brian Gaeke7283e7c2004-07-02 21:08:14 +00002044 implementation
2045
2046 int* %foo(%ST* %s) {
2047 entry:
2048 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002049 ret int* %reg
2050 }
2051</pre>
2052
Chris Lattner7faa8832002-04-14 06:13:44 +00002053<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002054
2055<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +00002056on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Chris Lattnere53e5082004-06-03 22:57:15 +00002057and <a href="#t_array">array</a> types require <tt>uint</tt>, <tt>int</tt>,
2058<tt>ulong</tt>, or <tt>long</tt> values, and <a href="#t_struct">structure</a>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002059types require <tt>uint</tt> <b>constants</b>.</p>
2060
Misha Brukman9d0919f2003-11-08 01:05:38 +00002061<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002062type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2063}</tt>' type, a structure. The second index indexes into the third element of
2064the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2065sbyte }</tt>' type, another structure. The third index indexes into the second
2066element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2067array. The two dimensions of the array are subscripted into, yielding an
John Criswellfc6b8952005-05-16 16:17:45 +00002068'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002069to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2070
Chris Lattner261efe92003-11-25 01:02:51 +00002071<p>Note that it is perfectly legal to index partially through a
2072structure, returning a pointer to an inner element. Because of this,
2073the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002074
2075<pre>
Chris Lattnerd4f6b172005-03-07 22:13:59 +00002076 int* %foo(%ST* %s) {
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002077 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2078 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2079 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2080 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2081 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2082 ret int* %t5
2083 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00002084</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00002085
2086<p>Note that it is undefined to access an array out of bounds: array and
2087pointer indexes must always be within the defined bounds of the array type.
2088The one exception for this rules is zero length arrays. These arrays are
2089defined to be accessible as variable length arrays, which requires access
2090beyond the zero'th element.</p>
2091
Chris Lattner7faa8832002-04-14 06:13:44 +00002092<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +00002093
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002094<pre>
2095 <i>; yields [12 x ubyte]*:aptr</i>
2096 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2097</pre>
2098
2099</div>
Chris Lattner00950542001-06-06 20:29:01 +00002100<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00002101<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002102<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +00002103<p>The instructions in this category are the "miscellaneous"
Chris Lattner261efe92003-11-25 01:02:51 +00002104instructions, which defy better classification.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002105</div>
Chris Lattner00950542001-06-06 20:29:01 +00002106<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002107<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
2108Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002109<div class="doc_text">
Chris Lattner33ba0d92001-07-09 00:26:23 +00002110<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002111<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002112<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002113<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
2114the SSA graph representing the function.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002115<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002116<p>The type of the incoming values are specified with the first type
2117field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
2118as arguments, with one pair for each predecessor basic block of the
2119current block. Only values of <a href="#t_firstclass">first class</a>
2120type may be used as the value arguments to the PHI node. Only labels
2121may be used as the label arguments.</p>
2122<p>There must be no non-phi instructions between the start of a basic
2123block and the PHI instructions: i.e. PHI instructions must be first in
2124a basic block.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002125<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002126<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
2127value specified by the parameter, depending on which basic block we
2128came from in the last <a href="#terminators">terminator</a> instruction.</p>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002129<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002130<pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi uint [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add uint %indvar, 1<br> br label %Loop<br></pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002131</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002132
Chris Lattner6536cfe2002-05-06 22:08:29 +00002133<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +00002134<div class="doc_subsubsection">
2135 <a name="i_cast">'<tt>cast .. to</tt>' Instruction</a>
2136</div>
2137
Misha Brukman9d0919f2003-11-08 01:05:38 +00002138<div class="doc_text">
Chris Lattnercc37aae2004-03-12 05:50:16 +00002139
Chris Lattner6536cfe2002-05-06 22:08:29 +00002140<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002141
2142<pre>
2143 &lt;result&gt; = cast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002144</pre>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002145
Chris Lattner6536cfe2002-05-06 22:08:29 +00002146<h5>Overview:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002147
2148<p>
2149The '<tt>cast</tt>' instruction is used as the primitive means to convert
2150integers to floating point, change data type sizes, and break type safety (by
2151casting pointers).
2152</p>
2153
2154
Chris Lattner6536cfe2002-05-06 22:08:29 +00002155<h5>Arguments:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002156
2157<p>
2158The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
2159class value, and a type to cast it to, which must also be a <a
2160href="#t_firstclass">first class</a> type.
2161</p>
2162
Chris Lattner6536cfe2002-05-06 22:08:29 +00002163<h5>Semantics:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002164
2165<p>
2166This instruction follows the C rules for explicit casts when determining how the
2167data being cast must change to fit in its new container.
2168</p>
2169
2170<p>
2171When casting to bool, any value that would be considered true in the context of
2172a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
2173all else are '<tt>false</tt>'.
2174</p>
2175
2176<p>
2177When extending an integral value from a type of one signness to another (for
2178example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
2179<b>source</b> value is signed, and zero-extended if the source value is
2180unsigned. <tt>bool</tt> values are always zero extended into either zero or
2181one.
2182</p>
2183
Chris Lattner33ba0d92001-07-09 00:26:23 +00002184<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002185
2186<pre>
2187 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
Chris Lattner7bae3952002-06-25 18:03:17 +00002188 %Y = cast int 123 to bool <i>; yields bool:true</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002189</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002190</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002191
2192<!-- _______________________________________________________________________ -->
2193<div class="doc_subsubsection">
2194 <a name="i_select">'<tt>select</tt>' Instruction</a>
2195</div>
2196
2197<div class="doc_text">
2198
2199<h5>Syntax:</h5>
2200
2201<pre>
2202 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
2203</pre>
2204
2205<h5>Overview:</h5>
2206
2207<p>
2208The '<tt>select</tt>' instruction is used to choose one value based on a
2209condition, without branching.
2210</p>
2211
2212
2213<h5>Arguments:</h5>
2214
2215<p>
2216The '<tt>select</tt>' instruction requires a boolean value indicating the condition, and two values of the same <a href="#t_firstclass">first class</a> type.
2217</p>
2218
2219<h5>Semantics:</h5>
2220
2221<p>
2222If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +00002223value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +00002224</p>
2225
2226<h5>Example:</h5>
2227
2228<pre>
2229 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
2230</pre>
2231</div>
2232
2233
Robert Bocchino3a558662006-01-05 17:37:02 +00002234<!-- _______________________________________________________________________ -->
2235<div class="doc_subsubsection">
2236 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2237</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002238
Robert Bocchino3a558662006-01-05 17:37:02 +00002239<div class="doc_text">
2240
2241<h5>Syntax:</h5>
2242
2243<pre>
2244 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, uint &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2245</pre>
2246
2247<h5>Overview:</h5>
2248
2249<p>
2250The '<tt>extractelement</tt>' instruction extracts a single scalar
2251element from a vector at a specified index.
2252</p>
2253
2254
2255<h5>Arguments:</h5>
2256
2257<p>
2258The first operand of an '<tt>extractelement</tt>' instruction is a
2259value of <a href="#t_packed">packed</a> type. The second operand is
2260an index indicating the position from which to extract the element.
2261The index may be a variable.</p>
2262
2263<h5>Semantics:</h5>
2264
2265<p>
2266The result is a scalar of the same type as the element type of
2267<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2268<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2269results are undefined.
2270</p>
2271
2272<h5>Example:</h5>
2273
2274<pre>
2275 %result = extractelement &lt;4 x int&gt; %vec, uint 0 <i>; yields int</i>
2276</pre>
2277</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002278
2279
Chris Lattner33ba0d92001-07-09 00:26:23 +00002280<!-- _______________________________________________________________________ -->
Chris Lattner2bff5242005-05-06 05:47:36 +00002281<div class="doc_subsubsection">
2282 <a name="i_call">'<tt>call</tt>' Instruction</a>
2283</div>
2284
Misha Brukman9d0919f2003-11-08 01:05:38 +00002285<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +00002286
Chris Lattner00950542001-06-06 20:29:01 +00002287<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002288<pre>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002289 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] &lt;ty&gt;* &lt;fnptrval&gt;(&lt;param list&gt;)
Chris Lattner2bff5242005-05-06 05:47:36 +00002290</pre>
2291
Chris Lattner00950542001-06-06 20:29:01 +00002292<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002293
Misha Brukman9d0919f2003-11-08 01:05:38 +00002294<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002295
Chris Lattner00950542001-06-06 20:29:01 +00002296<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002297
Misha Brukman9d0919f2003-11-08 01:05:38 +00002298<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002299
Chris Lattner6536cfe2002-05-06 22:08:29 +00002300<ol>
Chris Lattner261efe92003-11-25 01:02:51 +00002301 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002302 <p>The optional "tail" marker indicates whether the callee function accesses
2303 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +00002304 function call is eligible for tail call optimization. Note that calls may
2305 be marked "tail" even if they do not occur before a <a
2306 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +00002307 </li>
2308 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002309 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
2310 convention</a> the call should use. If none is specified, the call defaults
2311 to using C calling conventions.
2312 </li>
2313 <li>
Chris Lattner2bff5242005-05-06 05:47:36 +00002314 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
2315 being invoked. The argument types must match the types implied by this
John Criswellfc6b8952005-05-16 16:17:45 +00002316 signature. This type can be omitted if the function is not varargs and
2317 if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002318 </li>
2319 <li>
2320 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
2321 be invoked. In most cases, this is a direct function invocation, but
2322 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +00002323 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002324 </li>
2325 <li>
2326 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +00002327 function signature argument types. All arguments must be of
2328 <a href="#t_firstclass">first class</a> type. If the function signature
2329 indicates the function accepts a variable number of arguments, the extra
2330 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002331 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002332</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00002333
Chris Lattner00950542001-06-06 20:29:01 +00002334<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002335
Chris Lattner261efe92003-11-25 01:02:51 +00002336<p>The '<tt>call</tt>' instruction is used to cause control flow to
2337transfer to a specified function, with its incoming arguments bound to
2338the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
2339instruction in the called function, control flow continues with the
2340instruction after the function call, and the return value of the
2341function is bound to the result argument. This is a simpler case of
2342the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002343
Chris Lattner00950542001-06-06 20:29:01 +00002344<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002345
2346<pre>
2347 %retval = call int %test(int %argc)
2348 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
2349 %X = tail call int %foo()
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002350 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
Chris Lattner2bff5242005-05-06 05:47:36 +00002351</pre>
2352
Misha Brukman9d0919f2003-11-08 01:05:38 +00002353</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002354
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002355<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +00002356<div class="doc_subsubsection">
Chris Lattnere19d7a72004-09-27 21:51:25 +00002357 <a name="i_vaarg">'<tt>vaarg</tt>' Instruction</a>
2358</div>
2359
Misha Brukman9d0919f2003-11-08 01:05:38 +00002360<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +00002361
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002362<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002363
2364<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002365 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00002366</pre>
2367
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002368<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002369
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002370<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +00002371the "variable argument" area of a function call. It is used to implement the
2372<tt>va_arg</tt> macro in C.</p>
2373
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002374<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002375
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002376<p>This instruction takes a <tt>va_list*</tt> value and the type of
2377the argument. It returns a value of the specified argument type and
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00002378increments the <tt>va_list</tt> to point to the next argument. Again, the
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002379actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002380
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002381<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002382
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002383<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
2384type from the specified <tt>va_list</tt> and causes the
2385<tt>va_list</tt> to point to the next argument. For more information,
2386see the variable argument handling <a href="#int_varargs">Intrinsic
2387Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002388
2389<p>It is legal for this instruction to be called in a function which does not
2390take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002391function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002392
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002393<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +00002394href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +00002395argument.</p>
2396
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002397<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002398
2399<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
2400
Misha Brukman9d0919f2003-11-08 01:05:38 +00002401</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002402
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002403<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00002404<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
2405<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002406
Misha Brukman9d0919f2003-11-08 01:05:38 +00002407<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +00002408
2409<p>LLVM supports the notion of an "intrinsic function". These functions have
John Criswellfc6b8952005-05-16 16:17:45 +00002410well known names and semantics and are required to follow certain
Chris Lattner33aec9e2004-02-12 17:01:32 +00002411restrictions. Overall, these instructions represent an extension mechanism for
2412the LLVM language that does not require changing all of the transformations in
2413LLVM to add to the language (or the bytecode reader/writer, the parser,
2414etc...).</p>
2415
John Criswellfc6b8952005-05-16 16:17:45 +00002416<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
2417prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
Chris Lattner33aec9e2004-02-12 17:01:32 +00002418this. Intrinsic functions must always be external functions: you cannot define
2419the body of intrinsic functions. Intrinsic functions may only be used in call
2420or invoke instructions: it is illegal to take the address of an intrinsic
2421function. Additionally, because intrinsic functions are part of the LLVM
2422language, it is required that they all be documented here if any are added.</p>
2423
2424
John Criswellfc6b8952005-05-16 16:17:45 +00002425<p>To learn how to add an intrinsic function, please see the <a
Chris Lattner590cff32005-05-11 03:35:57 +00002426href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +00002427</p>
2428
Misha Brukman9d0919f2003-11-08 01:05:38 +00002429</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002430
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002431<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002432<div class="doc_subsection">
2433 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
2434</div>
2435
Misha Brukman9d0919f2003-11-08 01:05:38 +00002436<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002437
Misha Brukman9d0919f2003-11-08 01:05:38 +00002438<p>Variable argument support is defined in LLVM with the <a
Chris Lattner261efe92003-11-25 01:02:51 +00002439 href="#i_vanext"><tt>vanext</tt></a> instruction and these three
2440intrinsic functions. These functions are related to the similarly
2441named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002442
Chris Lattner261efe92003-11-25 01:02:51 +00002443<p>All of these functions operate on arguments that use a
2444target-specific value type "<tt>va_list</tt>". The LLVM assembly
2445language reference manual does not define what this type is, so all
2446transformations should be prepared to handle intrinsics with any type
2447used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002448
Misha Brukman9d0919f2003-11-08 01:05:38 +00002449<p>This example shows how the <a href="#i_vanext"><tt>vanext</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +00002450instruction and the variable argument handling intrinsic functions are
2451used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002452
Chris Lattner33aec9e2004-02-12 17:01:32 +00002453<pre>
2454int %test(int %X, ...) {
2455 ; Initialize variable argument processing
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002456 %ap = alloca sbyte*
2457 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002458
2459 ; Read a single integer argument
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002460 %tmp = va_arg sbyte** %ap, int
Chris Lattner33aec9e2004-02-12 17:01:32 +00002461
2462 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002463 %aq = alloca sbyte*
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002464 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002465 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002466
2467 ; Stop processing of arguments.
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002468 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002469 ret int %tmp
2470}
2471</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002472</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002473
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002474<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002475<div class="doc_subsubsection">
2476 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
2477</div>
2478
2479
Misha Brukman9d0919f2003-11-08 01:05:38 +00002480<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002481<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002482<pre> declare void %llvm.va_start(&lt;va_list&gt;* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002483<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002484<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
2485<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
2486href="#i_va_arg">va_arg</a></tt>.</p>
2487
2488<h5>Arguments:</h5>
2489
2490<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
2491
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002492<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002493
2494<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
2495macro available in C. In a target-dependent way, it initializes the
2496<tt>va_list</tt> element the argument points to, so that the next call to
2497<tt>va_arg</tt> will produce the first variable argument passed to the function.
2498Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
2499last argument of the function, the compiler can figure that out.</p>
2500
Misha Brukman9d0919f2003-11-08 01:05:38 +00002501</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002502
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002503<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002504<div class="doc_subsubsection">
2505 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
2506</div>
2507
Misha Brukman9d0919f2003-11-08 01:05:38 +00002508<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002509<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002510<pre> declare void %llvm.va_end(&lt;va_list*&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002511<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002512<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
2513which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
2514or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002515<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002516<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002517<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002518<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002519macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
2520Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
2521 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
2522with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002523</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002524
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002525<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002526<div class="doc_subsubsection">
2527 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
2528</div>
2529
Misha Brukman9d0919f2003-11-08 01:05:38 +00002530<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002531
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002532<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002533
2534<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002535 declare void %llvm.va_copy(&lt;va_list&gt;* &lt;destarglist&gt;,
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002536 &lt;va_list&gt;* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00002537</pre>
2538
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002539<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002540
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002541<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
2542the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002543
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002544<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002545
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002546<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002547The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002548
Chris Lattnerd7923912004-05-23 21:06:01 +00002549
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002550<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002551
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002552<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
2553available in C. In a target-dependent way, it copies the source
2554<tt>va_list</tt> element into the destination list. This intrinsic is necessary
2555because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
Chris Lattnerd7923912004-05-23 21:06:01 +00002556arbitrarily complex and require memory allocation, for example.</p>
2557
Misha Brukman9d0919f2003-11-08 01:05:38 +00002558</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002559
Chris Lattner33aec9e2004-02-12 17:01:32 +00002560<!-- ======================================================================= -->
2561<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +00002562 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
2563</div>
2564
2565<div class="doc_text">
2566
2567<p>
2568LLVM support for <a href="GarbageCollection.html">Accurate Garbage
2569Collection</a> requires the implementation and generation of these intrinsics.
2570These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
2571stack</a>, as well as garbage collector implementations that require <a
2572href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
2573Front-ends for type-safe garbage collected languages should generate these
2574intrinsics to make use of the LLVM garbage collectors. For more details, see <a
2575href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
2576</p>
2577</div>
2578
2579<!-- _______________________________________________________________________ -->
2580<div class="doc_subsubsection">
2581 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
2582</div>
2583
2584<div class="doc_text">
2585
2586<h5>Syntax:</h5>
2587
2588<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002589 declare void %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00002590</pre>
2591
2592<h5>Overview:</h5>
2593
John Criswell9e2485c2004-12-10 15:51:16 +00002594<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +00002595the code generator, and allows some metadata to be associated with it.</p>
2596
2597<h5>Arguments:</h5>
2598
2599<p>The first argument specifies the address of a stack object that contains the
2600root pointer. The second pointer (which must be either a constant or a global
2601value address) contains the meta-data to be associated with the root.</p>
2602
2603<h5>Semantics:</h5>
2604
2605<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
2606location. At compile-time, the code generator generates information to allow
2607the runtime to find the pointer at GC safe points.
2608</p>
2609
2610</div>
2611
2612
2613<!-- _______________________________________________________________________ -->
2614<div class="doc_subsubsection">
2615 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
2616</div>
2617
2618<div class="doc_text">
2619
2620<h5>Syntax:</h5>
2621
2622<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002623 declare sbyte* %llvm.gcread(sbyte** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00002624</pre>
2625
2626<h5>Overview:</h5>
2627
2628<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
2629locations, allowing garbage collector implementations that require read
2630barriers.</p>
2631
2632<h5>Arguments:</h5>
2633
2634<p>The argument is the address to read from, which should be an address
2635allocated from the garbage collector.</p>
2636
2637<h5>Semantics:</h5>
2638
2639<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
2640instruction, but may be replaced with substantially more complex code by the
2641garbage collector runtime, as needed.</p>
2642
2643</div>
2644
2645
2646<!-- _______________________________________________________________________ -->
2647<div class="doc_subsubsection">
2648 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
2649</div>
2650
2651<div class="doc_text">
2652
2653<h5>Syntax:</h5>
2654
2655<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002656 declare void %llvm.gcwrite(sbyte* %P1, sbyte** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00002657</pre>
2658
2659<h5>Overview:</h5>
2660
2661<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
2662locations, allowing garbage collector implementations that require write
2663barriers (such as generational or reference counting collectors).</p>
2664
2665<h5>Arguments:</h5>
2666
2667<p>The first argument is the reference to store, and the second is the heap
2668location to store to.</p>
2669
2670<h5>Semantics:</h5>
2671
2672<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
2673instruction, but may be replaced with substantially more complex code by the
2674garbage collector runtime, as needed.</p>
2675
2676</div>
2677
2678
2679
2680<!-- ======================================================================= -->
2681<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +00002682 <a name="int_codegen">Code Generator Intrinsics</a>
2683</div>
2684
2685<div class="doc_text">
2686<p>
2687These intrinsics are provided by LLVM to expose special features that may only
2688be implemented with code generator support.
2689</p>
2690
2691</div>
2692
2693<!-- _______________________________________________________________________ -->
2694<div class="doc_subsubsection">
2695 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
2696</div>
2697
2698<div class="doc_text">
2699
2700<h5>Syntax:</h5>
2701<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002702 declare void* %llvm.returnaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00002703</pre>
2704
2705<h5>Overview:</h5>
2706
2707<p>
2708The '<tt>llvm.returnaddress</tt>' intrinsic returns a target-specific value
2709indicating the return address of the current function or one of its callers.
2710</p>
2711
2712<h5>Arguments:</h5>
2713
2714<p>
2715The argument to this intrinsic indicates which function to return the address
2716for. Zero indicates the calling function, one indicates its caller, etc. The
2717argument is <b>required</b> to be a constant integer value.
2718</p>
2719
2720<h5>Semantics:</h5>
2721
2722<p>
2723The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
2724the return address of the specified call frame, or zero if it cannot be
2725identified. The value returned by this intrinsic is likely to be incorrect or 0
2726for arguments other than zero, so it should only be used for debugging purposes.
2727</p>
2728
2729<p>
2730Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00002731aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00002732source-language caller.
2733</p>
2734</div>
2735
2736
2737<!-- _______________________________________________________________________ -->
2738<div class="doc_subsubsection">
2739 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
2740</div>
2741
2742<div class="doc_text">
2743
2744<h5>Syntax:</h5>
2745<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002746 declare void* %llvm.frameaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00002747</pre>
2748
2749<h5>Overview:</h5>
2750
2751<p>
2752The '<tt>llvm.frameaddress</tt>' intrinsic returns the target-specific frame
2753pointer value for the specified stack frame.
2754</p>
2755
2756<h5>Arguments:</h5>
2757
2758<p>
2759The argument to this intrinsic indicates which function to return the frame
2760pointer for. Zero indicates the calling function, one indicates its caller,
2761etc. The argument is <b>required</b> to be a constant integer value.
2762</p>
2763
2764<h5>Semantics:</h5>
2765
2766<p>
2767The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
2768the frame address of the specified call frame, or zero if it cannot be
2769identified. The value returned by this intrinsic is likely to be incorrect or 0
2770for arguments other than zero, so it should only be used for debugging purposes.
2771</p>
2772
2773<p>
2774Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00002775aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00002776source-language caller.
2777</p>
2778</div>
2779
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002780<!-- _______________________________________________________________________ -->
2781<div class="doc_subsubsection">
2782 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
2783</div>
2784
2785<div class="doc_text">
2786
2787<h5>Syntax:</h5>
2788<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002789 declare void %llvm.prefetch(sbyte * &lt;address&gt;,
2790 uint &lt;rw&gt;, uint &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002791</pre>
2792
2793<h5>Overview:</h5>
2794
2795
2796<p>
2797The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +00002798a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
2799no
2800effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +00002801characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002802</p>
2803
2804<h5>Arguments:</h5>
2805
2806<p>
2807<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
2808determining if the fetch should be for a read (0) or write (1), and
2809<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +00002810locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002811<tt>locality</tt> arguments must be constant integers.
2812</p>
2813
2814<h5>Semantics:</h5>
2815
2816<p>
2817This intrinsic does not modify the behavior of the program. In particular,
2818prefetches cannot trap and do not produce a value. On targets that support this
2819intrinsic, the prefetch can provide hints to the processor cache for better
2820performance.
2821</p>
2822
2823</div>
2824
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002825<!-- _______________________________________________________________________ -->
2826<div class="doc_subsubsection">
2827 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
2828</div>
2829
2830<div class="doc_text">
2831
2832<h5>Syntax:</h5>
2833<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002834 declare void %llvm.pcmarker( uint &lt;id&gt; )
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002835</pre>
2836
2837<h5>Overview:</h5>
2838
2839
2840<p>
John Criswellfc6b8952005-05-16 16:17:45 +00002841The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
2842(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002843code to simulators and other tools. The method is target specific, but it is
2844expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00002845The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +00002846after optimizations. It is possible that the presence of a marker will inhibit
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002847optimizations. The intended use is to be inserted after optmizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +00002848correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002849</p>
2850
2851<h5>Arguments:</h5>
2852
2853<p>
2854<tt>id</tt> is a numerical id identifying the marker.
2855</p>
2856
2857<h5>Semantics:</h5>
2858
2859<p>
2860This intrinsic does not modify the behavior of the program. Backends that do not
2861support this intrinisic may ignore it.
2862</p>
2863
2864</div>
2865
Andrew Lenharth51b8d542005-11-11 16:47:30 +00002866<!-- _______________________________________________________________________ -->
2867<div class="doc_subsubsection">
2868 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
2869</div>
2870
2871<div class="doc_text">
2872
2873<h5>Syntax:</h5>
2874<pre>
2875 declare ulong %llvm.readcyclecounter( )
2876</pre>
2877
2878<h5>Overview:</h5>
2879
2880
2881<p>
2882The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
2883counter register (or similar low latency, high accuracy clocks) on those targets
2884that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
2885As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
2886should only be used for small timings.
2887</p>
2888
2889<h5>Semantics:</h5>
2890
2891<p>
2892When directly supported, reading the cycle counter should not modify any memory.
2893Implementations are allowed to either return a application specific value or a
2894system wide value. On backends without support, this is lowered to a constant 0.
2895</p>
2896
2897</div>
2898
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002899
John Criswell7123e272004-04-09 16:43:20 +00002900<!-- ======================================================================= -->
2901<div class="doc_subsection">
2902 <a name="int_os">Operating System Intrinsics</a>
2903</div>
2904
2905<div class="doc_text">
2906<p>
2907These intrinsics are provided by LLVM to support the implementation of
2908operating system level code.
2909</p>
2910
2911</div>
John Criswell183402a2004-04-12 15:02:16 +00002912
John Criswellcfd3bac2004-04-09 15:23:37 +00002913<!-- _______________________________________________________________________ -->
2914<div class="doc_subsubsection">
2915 <a name="i_readport">'<tt>llvm.readport</tt>' Intrinsic</a>
2916</div>
2917
2918<div class="doc_text">
2919
2920<h5>Syntax:</h5>
2921<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002922 declare &lt;integer type&gt; %llvm.readport (&lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +00002923</pre>
2924
2925<h5>Overview:</h5>
2926
2927<p>
John Criswell7123e272004-04-09 16:43:20 +00002928The '<tt>llvm.readport</tt>' intrinsic reads data from the specified hardware
2929I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +00002930</p>
2931
2932<h5>Arguments:</h5>
2933
2934<p>
John Criswell7123e272004-04-09 16:43:20 +00002935The argument to this intrinsic indicates the hardware I/O address from which
2936to read the data. The address is in the hardware I/O address namespace (as
2937opposed to being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +00002938</p>
2939
2940<h5>Semantics:</h5>
2941
2942<p>
John Criswell7123e272004-04-09 16:43:20 +00002943The '<tt>llvm.readport</tt>' intrinsic reads data from the hardware I/O port
2944specified by <i>address</i> and returns the value. The address and return
2945value must be integers, but the size is dependent upon the platform upon which
2946the program is code generated. For example, on x86, the address must be an
Misha Brukmancfa87bc2005-04-22 18:02:52 +00002947unsigned 16-bit value, and the return value must be 8, 16, or 32 bits.
John Criswellcfd3bac2004-04-09 15:23:37 +00002948</p>
2949
2950</div>
2951
2952<!-- _______________________________________________________________________ -->
2953<div class="doc_subsubsection">
2954 <a name="i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a>
2955</div>
2956
2957<div class="doc_text">
2958
2959<h5>Syntax:</h5>
2960<pre>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002961 call void (&lt;integer type&gt;, &lt;integer type&gt;)*
2962 %llvm.writeport (&lt;integer type&gt; &lt;value&gt;,
2963 &lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +00002964</pre>
2965
2966<h5>Overview:</h5>
2967
2968<p>
John Criswell7123e272004-04-09 16:43:20 +00002969The '<tt>llvm.writeport</tt>' intrinsic writes data to the specified hardware
2970I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +00002971</p>
2972
2973<h5>Arguments:</h5>
2974
2975<p>
John Criswell96db6fc2004-04-12 16:33:19 +00002976The first argument is the value to write to the I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +00002977</p>
2978
2979<p>
John Criswell96db6fc2004-04-12 16:33:19 +00002980The second argument indicates the hardware I/O address to which data should be
2981written. The address is in the hardware I/O address namespace (as opposed to
2982being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +00002983</p>
2984
2985<h5>Semantics:</h5>
2986
2987<p>
2988The '<tt>llvm.writeport</tt>' intrinsic writes <i>value</i> to the I/O port
2989specified by <i>address</i>. The address and value must be integers, but the
2990size is dependent upon the platform upon which the program is code generated.
Misha Brukmancfa87bc2005-04-22 18:02:52 +00002991For example, on x86, the address must be an unsigned 16-bit value, and the
John Criswell7123e272004-04-09 16:43:20 +00002992value written must be 8, 16, or 32 bits in length.
John Criswellcfd3bac2004-04-09 15:23:37 +00002993</p>
2994
2995</div>
Chris Lattner10610642004-02-14 04:08:35 +00002996
John Criswell183402a2004-04-12 15:02:16 +00002997<!-- _______________________________________________________________________ -->
2998<div class="doc_subsubsection">
2999 <a name="i_readio">'<tt>llvm.readio</tt>' Intrinsic</a>
3000</div>
3001
3002<div class="doc_text">
3003
3004<h5>Syntax:</h5>
3005<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003006 declare &lt;result&gt; %llvm.readio (&lt;ty&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +00003007</pre>
3008
3009<h5>Overview:</h5>
3010
3011<p>
3012The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
3013address.
3014</p>
3015
3016<h5>Arguments:</h5>
3017
3018<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003019The argument to this intrinsic is a pointer indicating the memory address from
3020which to read the data. The data must be a
3021<a href="#t_firstclass">first class</a> type.
John Criswell183402a2004-04-12 15:02:16 +00003022</p>
3023
3024<h5>Semantics:</h5>
3025
3026<p>
3027The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
John Criswell96db6fc2004-04-12 16:33:19 +00003028location specified by <i>pointer</i> and returns the value. The argument must
3029be a pointer, and the return value must be a
3030<a href="#t_firstclass">first class</a> type. However, certain architectures
Misha Brukmancfa87bc2005-04-22 18:02:52 +00003031may not support I/O on all first class types. For example, 32-bit processors
John Criswell96db6fc2004-04-12 16:33:19 +00003032may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +00003033</p>
3034
3035<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003036This intrinsic enforces an in-order memory model for llvm.readio and
3037llvm.writeio calls on machines that use dynamic scheduling. Dynamically
3038scheduled processors may execute loads and stores out of order, re-ordering at
3039run time accesses to memory mapped I/O registers. Using these intrinsics
3040ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +00003041</p>
3042
3043</div>
3044
3045<!-- _______________________________________________________________________ -->
3046<div class="doc_subsubsection">
3047 <a name="i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a>
3048</div>
3049
3050<div class="doc_text">
3051
3052<h5>Syntax:</h5>
3053<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003054 declare void %llvm.writeio (&lt;ty1&gt; &lt;value&gt;, &lt;ty2&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +00003055</pre>
3056
3057<h5>Overview:</h5>
3058
3059<p>
3060The '<tt>llvm.writeio</tt>' intrinsic writes data to the specified memory
3061mapped I/O address.
3062</p>
3063
3064<h5>Arguments:</h5>
3065
3066<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003067The first argument is the value to write to the memory mapped I/O location.
3068The second argument is a pointer indicating the memory address to which the
3069data should be written.
John Criswell183402a2004-04-12 15:02:16 +00003070</p>
3071
3072<h5>Semantics:</h5>
3073
3074<p>
3075The '<tt>llvm.writeio</tt>' intrinsic writes <i>value</i> to the memory mapped
John Criswell96db6fc2004-04-12 16:33:19 +00003076I/O address specified by <i>pointer</i>. The value must be a
3077<a href="#t_firstclass">first class</a> type. However, certain architectures
Misha Brukmancfa87bc2005-04-22 18:02:52 +00003078may not support I/O on all first class types. For example, 32-bit processors
John Criswell96db6fc2004-04-12 16:33:19 +00003079may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +00003080</p>
3081
3082<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003083This intrinsic enforces an in-order memory model for llvm.readio and
3084llvm.writeio calls on machines that use dynamic scheduling. Dynamically
3085scheduled processors may execute loads and stores out of order, re-ordering at
3086run time accesses to memory mapped I/O registers. Using these intrinsics
3087ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +00003088</p>
3089
3090</div>
3091
Chris Lattner10610642004-02-14 04:08:35 +00003092<!-- ======================================================================= -->
3093<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +00003094 <a name="int_libc">Standard C Library Intrinsics</a>
3095</div>
3096
3097<div class="doc_text">
3098<p>
Chris Lattner10610642004-02-14 04:08:35 +00003099LLVM provides intrinsics for a few important standard C library functions.
3100These intrinsics allow source-language front-ends to pass information about the
3101alignment of the pointer arguments to the code generator, providing opportunity
3102for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003103</p>
3104
3105</div>
3106
3107<!-- _______________________________________________________________________ -->
3108<div class="doc_subsubsection">
3109 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
3110</div>
3111
3112<div class="doc_text">
3113
3114<h5>Syntax:</h5>
3115<pre>
Reid Spencerd4622352005-04-26 20:41:16 +00003116 declare void %llvm.memcpy(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3117 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003118</pre>
3119
3120<h5>Overview:</h5>
3121
3122<p>
3123The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
3124location to the destination location.
3125</p>
3126
3127<p>
3128Note that, unlike the standard libc function, the <tt>llvm.memcpy</tt> intrinsic
3129does not return a value, and takes an extra alignment argument.
3130</p>
3131
3132<h5>Arguments:</h5>
3133
3134<p>
3135The first argument is a pointer to the destination, the second is a pointer to
3136the source. The third argument is an (arbitrarily sized) integer argument
3137specifying the number of bytes to copy, and the fourth argument is the alignment
3138of the source and destination locations.
3139</p>
3140
Chris Lattner3301ced2004-02-12 21:18:15 +00003141<p>
3142If the call to this intrinisic has an alignment value that is not 0 or 1, then
3143the caller guarantees that the size of the copy is a multiple of the alignment
3144and that both the source and destination pointers are aligned to that boundary.
3145</p>
3146
Chris Lattner33aec9e2004-02-12 17:01:32 +00003147<h5>Semantics:</h5>
3148
3149<p>
3150The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
3151location to the destination location, which are not allowed to overlap. It
3152copies "len" bytes of memory over. If the argument is known to be aligned to
3153some boundary, this can be specified as the fourth argument, otherwise it should
3154be set to 0 or 1.
3155</p>
3156</div>
3157
3158
Chris Lattner0eb51b42004-02-12 18:10:10 +00003159<!-- _______________________________________________________________________ -->
3160<div class="doc_subsubsection">
3161 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
3162</div>
3163
3164<div class="doc_text">
3165
3166<h5>Syntax:</h5>
3167<pre>
Reid Spencerd4622352005-04-26 20:41:16 +00003168 declare void %llvm.memmove(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3169 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00003170</pre>
3171
3172<h5>Overview:</h5>
3173
3174<p>
3175The '<tt>llvm.memmove</tt>' intrinsic moves a block of memory from the source
3176location to the destination location. It is similar to the '<tt>llvm.memcpy</tt>'
3177intrinsic but allows the two memory locations to overlap.
3178</p>
3179
3180<p>
3181Note that, unlike the standard libc function, the <tt>llvm.memmove</tt> intrinsic
3182does not return a value, and takes an extra alignment argument.
3183</p>
3184
3185<h5>Arguments:</h5>
3186
3187<p>
3188The first argument is a pointer to the destination, the second is a pointer to
3189the source. The third argument is an (arbitrarily sized) integer argument
3190specifying the number of bytes to copy, and the fourth argument is the alignment
3191of the source and destination locations.
3192</p>
3193
Chris Lattner3301ced2004-02-12 21:18:15 +00003194<p>
3195If the call to this intrinisic has an alignment value that is not 0 or 1, then
3196the caller guarantees that the size of the copy is a multiple of the alignment
3197and that both the source and destination pointers are aligned to that boundary.
3198</p>
3199
Chris Lattner0eb51b42004-02-12 18:10:10 +00003200<h5>Semantics:</h5>
3201
3202<p>
3203The '<tt>llvm.memmove</tt>' intrinsic copies a block of memory from the source
3204location to the destination location, which may overlap. It
3205copies "len" bytes of memory over. If the argument is known to be aligned to
3206some boundary, this can be specified as the fourth argument, otherwise it should
3207be set to 0 or 1.
3208</p>
3209</div>
3210
Chris Lattner8ff75902004-01-06 05:31:32 +00003211
Chris Lattner10610642004-02-14 04:08:35 +00003212<!-- _______________________________________________________________________ -->
3213<div class="doc_subsubsection">
3214 <a name="i_memset">'<tt>llvm.memset</tt>' Intrinsic</a>
3215</div>
3216
3217<div class="doc_text">
3218
3219<h5>Syntax:</h5>
3220<pre>
Reid Spencerd4622352005-04-26 20:41:16 +00003221 declare void %llvm.memset(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
3222 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003223</pre>
3224
3225<h5>Overview:</h5>
3226
3227<p>
3228The '<tt>llvm.memset</tt>' intrinsic fills a block of memory with a particular
3229byte value.
3230</p>
3231
3232<p>
3233Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
3234does not return a value, and takes an extra alignment argument.
3235</p>
3236
3237<h5>Arguments:</h5>
3238
3239<p>
3240The first argument is a pointer to the destination to fill, the second is the
3241byte value to fill it with, the third argument is an (arbitrarily sized) integer
3242argument specifying the number of bytes to fill, and the fourth argument is the
3243known alignment of destination location.
3244</p>
3245
3246<p>
3247If the call to this intrinisic has an alignment value that is not 0 or 1, then
3248the caller guarantees that the size of the copy is a multiple of the alignment
3249and that the destination pointer is aligned to that boundary.
3250</p>
3251
3252<h5>Semantics:</h5>
3253
3254<p>
3255The '<tt>llvm.memset</tt>' intrinsic fills "len" bytes of memory starting at the
3256destination location. If the argument is known to be aligned to some boundary,
3257this can be specified as the fourth argument, otherwise it should be set to 0 or
32581.
3259</p>
3260</div>
3261
3262
Chris Lattner32006282004-06-11 02:28:03 +00003263<!-- _______________________________________________________________________ -->
3264<div class="doc_subsubsection">
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003265 <a name="i_isunordered">'<tt>llvm.isunordered</tt>' Intrinsic</a>
3266</div>
3267
3268<div class="doc_text">
3269
3270<h5>Syntax:</h5>
3271<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003272 declare bool %llvm.isunordered(&lt;float or double&gt; Val1, &lt;float or double&gt; Val2)
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003273</pre>
3274
3275<h5>Overview:</h5>
3276
3277<p>
3278The '<tt>llvm.isunordered</tt>' intrinsic returns true if either or both of the
3279specified floating point values is a NAN.
3280</p>
3281
3282<h5>Arguments:</h5>
3283
3284<p>
3285The arguments are floating point numbers of the same type.
3286</p>
3287
3288<h5>Semantics:</h5>
3289
3290<p>
3291If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
3292false.
3293</p>
3294</div>
3295
3296
Chris Lattnera4d74142005-07-21 01:29:16 +00003297<!-- _______________________________________________________________________ -->
3298<div class="doc_subsubsection">
3299 <a name="i_sqrt">'<tt>llvm.sqrt</tt>' Intrinsic</a>
3300</div>
3301
3302<div class="doc_text">
3303
3304<h5>Syntax:</h5>
3305<pre>
3306 declare &lt;float or double&gt; %llvm.sqrt(&lt;float or double&gt; Val)
3307</pre>
3308
3309<h5>Overview:</h5>
3310
3311<p>
3312The '<tt>llvm.sqrt</tt>' intrinsic returns the sqrt of the specified operand,
3313returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
3314<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
3315negative numbers (which allows for better optimization).
3316</p>
3317
3318<h5>Arguments:</h5>
3319
3320<p>
3321The argument and return value are floating point numbers of the same type.
3322</p>
3323
3324<h5>Semantics:</h5>
3325
3326<p>
3327This function returns the sqrt of the specified operand if it is a positive
3328floating point number.
3329</p>
3330</div>
3331
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003332<!-- ======================================================================= -->
3333<div class="doc_subsection">
3334 <a name="int_count">Bit Counting Intrinsics</a>
3335</div>
3336
3337<div class="doc_text">
3338<p>
3339LLVM provides intrinsics for a few important bit counting operations.
3340These allow efficient code generation for some algorithms.
3341</p>
3342
3343</div>
3344
3345<!-- _______________________________________________________________________ -->
3346<div class="doc_subsubsection">
3347 <a name="int_ctpop">'<tt>llvm.ctpop</tt>' Intrinsic</a>
3348</div>
3349
3350<div class="doc_text">
3351
3352<h5>Syntax:</h5>
3353<pre>
3354 declare int %llvm.ctpop(int &lt;src&gt;)
3355
3356</pre>
3357
3358<h5>Overview:</h5>
3359
3360<p>
3361The '<tt>llvm.ctpop</tt>' intrinsic counts the number of ones in a variable.
3362</p>
3363
3364<h5>Arguments:</h5>
3365
3366<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003367The only argument is the value to be counted. The argument may be of any
3368integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003369</p>
3370
3371<h5>Semantics:</h5>
3372
3373<p>
3374The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
3375</p>
3376</div>
3377
3378<!-- _______________________________________________________________________ -->
3379<div class="doc_subsubsection">
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003380 <a name="int_ctlz">'<tt>llvm.ctlz</tt>' Intrinsic</a>
3381</div>
3382
3383<div class="doc_text">
3384
3385<h5>Syntax:</h5>
3386<pre>
3387 declare int %llvm.ctlz(int &lt;src&gt;)
3388
3389</pre>
3390
3391<h5>Overview:</h5>
3392
3393<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003394The '<tt>llvm.ctlz</tt>' intrinsic counts the number of leading zeros in a
3395variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003396</p>
3397
3398<h5>Arguments:</h5>
3399
3400<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003401The only argument is the value to be counted. The argument may be of any
3402integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003403</p>
3404
3405<h5>Semantics:</h5>
3406
3407<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00003408The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
3409in a variable. If the src == 0 then the result is the size in bits of the type
3410of src. For example, <tt>llvm.cttz(int 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003411</p>
3412</div>
Chris Lattner32006282004-06-11 02:28:03 +00003413
3414
Chris Lattnereff29ab2005-05-15 19:39:26 +00003415
3416<!-- _______________________________________________________________________ -->
3417<div class="doc_subsubsection">
3418 <a name="int_cttz">'<tt>llvm.cttz</tt>' Intrinsic</a>
3419</div>
3420
3421<div class="doc_text">
3422
3423<h5>Syntax:</h5>
3424<pre>
3425 declare int %llvm.cttz(int &lt;src&gt;)
3426
3427</pre>
3428
3429<h5>Overview:</h5>
3430
3431<p>
3432The '<tt>llvm.cttz</tt>' intrinsic counts the number of trailing zeros.
3433</p>
3434
3435<h5>Arguments:</h5>
3436
3437<p>
3438The only argument is the value to be counted. The argument may be of any
3439integer type. The return type must match the argument type.
3440</p>
3441
3442<h5>Semantics:</h5>
3443
3444<p>
3445The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
3446in a variable. If the src == 0 then the result is the size in bits of the type
3447of src. For example, <tt>llvm.cttz(2) = 1</tt>.
3448</p>
3449</div>
3450
Chris Lattner8ff75902004-01-06 05:31:32 +00003451<!-- ======================================================================= -->
3452<div class="doc_subsection">
3453 <a name="int_debugger">Debugger Intrinsics</a>
3454</div>
3455
3456<div class="doc_text">
3457<p>
3458The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
3459are described in the <a
3460href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
3461Debugging</a> document.
3462</p>
3463</div>
3464
3465
Chris Lattner00950542001-06-06 20:29:01 +00003466<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00003467<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00003468<address>
3469 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3470 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3471 <a href="http://validator.w3.org/check/referer"><img
3472 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3473
3474 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
3475 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
3476 Last modified: $Date$
3477</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003478</body>
3479</html>