blob: 751efa1b855b5736f2f61f9105ee0dfacba1966b [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 Lattner57e1f392006-01-13 02:03:13 +0000130 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
131 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000132 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000133 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000134 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000135 </ol>
136 </li>
137 <li><a href="#int_os">Operating System Intrinsics</a>
138 <ol>
Chris Lattner32006282004-06-11 02:28:03 +0000139 <li><a href="#i_readport">'<tt>llvm.readport</tt>' Intrinsic</a></li>
140 <li><a href="#i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a></li>
John Criswell183402a2004-04-12 15:02:16 +0000141 <li><a href="#i_readio">'<tt>llvm.readio</tt>' Intrinsic</a></li>
142 <li><a href="#i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000143 </ol>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000144 <li><a href="#int_libc">Standard C Library Intrinsics</a>
145 <ol>
146 <li><a href="#i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a></li>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000147 <li><a href="#i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000148 <li><a href="#i_memset">'<tt>llvm.memset</tt>' Intrinsic</a></li>
Alkis Evlogimenos96853722004-06-12 19:19:14 +0000149 <li><a href="#i_isunordered">'<tt>llvm.isunordered</tt>' Intrinsic</a></li>
Chris Lattnera4d74142005-07-21 01:29:16 +0000150 <li><a href="#i_sqrt">'<tt>llvm.sqrt</tt>' Intrinsic</a></li>
151
Chris Lattner33aec9e2004-02-12 17:01:32 +0000152 </ol>
153 </li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000154 <li><a href="#int_count">Bit counting Intrinsics</a>
155 <ol>
156 <li><a href="#int_ctpop">'<tt>llvm.ctpop</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000157 <li><a href="#int_ctlz">'<tt>llvm.ctlz</tt>' Intrinsic </a></li>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000158 <li><a href="#int_cttz">'<tt>llvm.cttz</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000159 </ol>
160 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000161 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000162 </ol>
163 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000164</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000165
166<div class="doc_author">
167 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
168 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000169</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000170
Chris Lattner00950542001-06-06 20:29:01 +0000171<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000172<div class="doc_section"> <a name="abstract">Abstract </a></div>
173<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000174
Misha Brukman9d0919f2003-11-08 01:05:38 +0000175<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000176<p>This document is a reference manual for the LLVM assembly language.
177LLVM is an SSA based representation that provides type safety,
178low-level operations, flexibility, and the capability of representing
179'all' high-level languages cleanly. It is the common code
180representation used throughout all phases of the LLVM compilation
181strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000182</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000183
Chris Lattner00950542001-06-06 20:29:01 +0000184<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000185<div class="doc_section"> <a name="introduction">Introduction</a> </div>
186<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000187
Misha Brukman9d0919f2003-11-08 01:05:38 +0000188<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000189
Chris Lattner261efe92003-11-25 01:02:51 +0000190<p>The LLVM code representation is designed to be used in three
191different forms: as an in-memory compiler IR, as an on-disk bytecode
192representation (suitable for fast loading by a Just-In-Time compiler),
193and as a human readable assembly language representation. This allows
194LLVM to provide a powerful intermediate representation for efficient
195compiler transformations and analysis, while providing a natural means
196to debug and visualize the transformations. The three different forms
197of LLVM are all equivalent. This document describes the human readable
198representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000199
John Criswellc1f786c2005-05-13 22:25:59 +0000200<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000201while being expressive, typed, and extensible at the same time. It
202aims to be a "universal IR" of sorts, by being at a low enough level
203that high-level ideas may be cleanly mapped to it (similar to how
204microprocessors are "universal IR's", allowing many source languages to
205be mapped to them). By providing type information, LLVM can be used as
206the target of optimizations: for example, through pointer analysis, it
207can be proven that a C automatic variable is never accessed outside of
208the current function... allowing it to be promoted to a simple SSA
209value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000210
Misha Brukman9d0919f2003-11-08 01:05:38 +0000211</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000212
Chris Lattner00950542001-06-06 20:29:01 +0000213<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000214<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000215
Misha Brukman9d0919f2003-11-08 01:05:38 +0000216<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000217
Chris Lattner261efe92003-11-25 01:02:51 +0000218<p>It is important to note that this document describes 'well formed'
219LLVM assembly language. There is a difference between what the parser
220accepts and what is considered 'well formed'. For example, the
221following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000222
223<pre>
224 %x = <a href="#i_add">add</a> int 1, %x
225</pre>
226
Chris Lattner261efe92003-11-25 01:02:51 +0000227<p>...because the definition of <tt>%x</tt> does not dominate all of
228its uses. The LLVM infrastructure provides a verification pass that may
229be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000230automatically run by the parser after parsing input assembly and by
Chris Lattner261efe92003-11-25 01:02:51 +0000231the optimizer before it outputs bytecode. The violations pointed out
232by the verifier pass indicate bugs in transformation passes or input to
233the parser.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000234
Chris Lattner261efe92003-11-25 01:02:51 +0000235<!-- Describe the typesetting conventions here. --> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000236
Chris Lattner00950542001-06-06 20:29:01 +0000237<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000238<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000239<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000240
Misha Brukman9d0919f2003-11-08 01:05:38 +0000241<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000242
Chris Lattner261efe92003-11-25 01:02:51 +0000243<p>LLVM uses three different forms of identifiers, for different
244purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000245
Chris Lattner00950542001-06-06 20:29:01 +0000246<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000247 <li>Named values are represented as a string of characters with a '%' prefix.
248 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
249 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
250 Identifiers which require other characters in their names can be surrounded
251 with quotes. In this way, anything except a <tt>"</tt> character can be used
252 in a name.</li>
253
254 <li>Unnamed values are represented as an unsigned numeric value with a '%'
255 prefix. For example, %12, %2, %44.</li>
256
Reid Spencercc16dc32004-12-09 18:02:53 +0000257 <li>Constants, which are described in a <a href="#constants">section about
258 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000259</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000260
261<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
262don't need to worry about name clashes with reserved words, and the set of
263reserved words may be expanded in the future without penalty. Additionally,
264unnamed identifiers allow a compiler to quickly come up with a temporary
265variable without having to avoid symbol table conflicts.</p>
266
Chris Lattner261efe92003-11-25 01:02:51 +0000267<p>Reserved words in LLVM are very similar to reserved words in other
268languages. There are keywords for different opcodes ('<tt><a
Chris Lattnere5d947b2004-12-09 16:36:40 +0000269href="#i_add">add</a></tt>', '<tt><a href="#i_cast">cast</a></tt>', '<tt><a
270href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
271href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
272and others. These reserved words cannot conflict with variable names, because
273none of them start with a '%' character.</p>
274
275<p>Here is an example of LLVM code to multiply the integer variable
276'<tt>%X</tt>' by 8:</p>
277
Misha Brukman9d0919f2003-11-08 01:05:38 +0000278<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000279
280<pre>
281 %result = <a href="#i_mul">mul</a> uint %X, 8
282</pre>
283
Misha Brukman9d0919f2003-11-08 01:05:38 +0000284<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000285
286<pre>
287 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
288</pre>
289
Misha Brukman9d0919f2003-11-08 01:05:38 +0000290<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000291
292<pre>
293 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
294 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
295 %result = <a href="#i_add">add</a> uint %1, %1
296</pre>
297
Chris Lattner261efe92003-11-25 01:02:51 +0000298<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
299important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000300
Chris Lattner00950542001-06-06 20:29:01 +0000301<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000302
303 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
304 line.</li>
305
306 <li>Unnamed temporaries are created when the result of a computation is not
307 assigned to a named value.</li>
308
Misha Brukman9d0919f2003-11-08 01:05:38 +0000309 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000310
Misha Brukman9d0919f2003-11-08 01:05:38 +0000311</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000312
John Criswelle4c57cc2005-05-12 16:52:32 +0000313<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000314demonstrating instructions, we will follow an instruction with a comment that
315defines the type and name of value produced. Comments are shown in italic
316text.</p>
317
Misha Brukman9d0919f2003-11-08 01:05:38 +0000318</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000319
320<!-- *********************************************************************** -->
321<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
322<!-- *********************************************************************** -->
323
324<!-- ======================================================================= -->
325<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
326</div>
327
328<div class="doc_text">
329
330<p>LLVM programs are composed of "Module"s, each of which is a
331translation unit of the input programs. Each module consists of
332functions, global variables, and symbol table entries. Modules may be
333combined together with the LLVM linker, which merges function (and
334global variable) definitions, resolves forward declarations, and merges
335symbol table entries. Here is an example of the "hello world" module:</p>
336
337<pre><i>; Declare the string constant as a global constant...</i>
338<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
339 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
340
341<i>; External declaration of the puts function</i>
342<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
343
344<i>; Definition of main function</i>
345int %main() { <i>; int()* </i>
346 <i>; Convert [13x sbyte]* to sbyte *...</i>
347 %cast210 = <a
348 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
349
350 <i>; Call puts function to write out the string to stdout...</i>
351 <a
352 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
353 <a
354 href="#i_ret">ret</a> int 0<br>}<br></pre>
355
356<p>This example is made up of a <a href="#globalvars">global variable</a>
357named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
358function, and a <a href="#functionstructure">function definition</a>
359for "<tt>main</tt>".</p>
360
Chris Lattnere5d947b2004-12-09 16:36:40 +0000361<p>In general, a module is made up of a list of global values,
362where both functions and global variables are global values. Global values are
363represented by a pointer to a memory location (in this case, a pointer to an
364array of char, and a pointer to a function), and have one of the following <a
365href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000366
Chris Lattnere5d947b2004-12-09 16:36:40 +0000367</div>
368
369<!-- ======================================================================= -->
370<div class="doc_subsection">
371 <a name="linkage">Linkage Types</a>
372</div>
373
374<div class="doc_text">
375
376<p>
377All Global Variables and Functions have one of the following types of linkage:
378</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000379
380<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000381
Chris Lattnerfa730212004-12-09 16:11:40 +0000382 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000383
384 <dd>Global values with internal linkage are only directly accessible by
385 objects in the current module. In particular, linking code into a module with
386 an internal global value may cause the internal to be renamed as necessary to
387 avoid collisions. Because the symbol is internal to the module, all
388 references can be updated. This corresponds to the notion of the
389 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattnerfa730212004-12-09 16:11:40 +0000390 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000391
Chris Lattnerfa730212004-12-09 16:11:40 +0000392 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000393
394 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
395 the twist that linking together two modules defining the same
396 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
397 is typically used to implement inline functions. Unreferenced
398 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000399 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000400
Chris Lattnerfa730212004-12-09 16:11:40 +0000401 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000402
403 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
404 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
405 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000406 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000407
Chris Lattnerfa730212004-12-09 16:11:40 +0000408 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000409
410 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
411 pointer to array type. When two global variables with appending linkage are
412 linked together, the two global arrays are appended together. This is the
413 LLVM, typesafe, equivalent of having the system linker append together
414 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000415 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000416
Chris Lattnerfa730212004-12-09 16:11:40 +0000417 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000418
419 <dd>If none of the above identifiers are used, the global is externally
420 visible, meaning that it participates in linkage and can be used to resolve
421 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000422 </dd>
423</dl>
424
Chris Lattnerfa730212004-12-09 16:11:40 +0000425<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
426variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
427variable and was linked with this one, one of the two would be renamed,
428preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
429external (i.e., lacking any linkage declarations), they are accessible
430outside of the current module. It is illegal for a function <i>declaration</i>
431to have any linkage type other than "externally visible".</a></p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000432
Chris Lattnerfa730212004-12-09 16:11:40 +0000433</div>
434
435<!-- ======================================================================= -->
436<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000437 <a name="callingconv">Calling Conventions</a>
438</div>
439
440<div class="doc_text">
441
442<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
443and <a href="#i_invoke">invokes</a> can all have an optional calling convention
444specified for the call. The calling convention of any pair of dynamic
445caller/callee must match, or the behavior of the program is undefined. The
446following calling conventions are supported by LLVM, and more may be added in
447the future:</p>
448
449<dl>
450 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
451
452 <dd>This calling convention (the default if no other calling convention is
453 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000454 supports varargs function calls and tolerates some mismatch in the declared
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000455 prototype and implemented declaration of the function (as does normal C).
456 </dd>
457
458 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
459
460 <dd>This calling convention attempts to make calls as fast as possible
461 (e.g. by passing things in registers). This calling convention allows the
462 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000463 without having to conform to an externally specified ABI. Implementations of
464 this convention should allow arbitrary tail call optimization to be supported.
465 This calling convention does not support varargs and requires the prototype of
466 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000467 </dd>
468
469 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
470
471 <dd>This calling convention attempts to make code in the caller as efficient
472 as possible under the assumption that the call is not commonly executed. As
473 such, these calls often preserve all registers so that the call does not break
474 any live ranges in the caller side. This calling convention does not support
475 varargs and requires the prototype of all callees to exactly match the
476 prototype of the function definition.
477 </dd>
478
Chris Lattnercfe6b372005-05-07 01:46:40 +0000479 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000480
481 <dd>Any calling convention may be specified by number, allowing
482 target-specific calling conventions to be used. Target specific calling
483 conventions start at 64.
484 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000485</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000486
487<p>More calling conventions can be added/defined on an as-needed basis, to
488support pascal conventions or any other well-known target-independent
489convention.</p>
490
491</div>
492
493<!-- ======================================================================= -->
494<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000495 <a name="globalvars">Global Variables</a>
496</div>
497
498<div class="doc_text">
499
Chris Lattner3689a342005-02-12 19:30:21 +0000500<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000501instead of run-time. Global variables may optionally be initialized, may have
502an explicit section to be placed in, and may
Chris Lattner2cbdc452005-11-06 08:02:57 +0000503have an optional explicit alignment specified. A
John Criswell0ec250c2005-10-24 16:17:18 +0000504variable may be defined as a global "constant," which indicates that the
Chris Lattner3689a342005-02-12 19:30:21 +0000505contents of the variable will <b>never</b> be modified (enabling better
506optimization, allowing the global data to be placed in the read-only section of
507an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000508cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000509
510<p>
511LLVM explicitly allows <em>declarations</em> of global variables to be marked
512constant, even if the final definition of the global is not. This capability
513can be used to enable slightly better optimization of the program, but requires
514the language definition to guarantee that optimizations based on the
515'constantness' are valid for the translation units that do not include the
516definition.
517</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000518
519<p>As SSA values, global variables define pointer values that are in
520scope (i.e. they dominate) all basic blocks in the program. Global
521variables always define a pointer to their "content" type because they
522describe a region of memory, and all memory objects in LLVM are
523accessed through pointers.</p>
524
Chris Lattner88f6c462005-11-12 00:45:07 +0000525<p>LLVM allows an explicit section to be specified for globals. If the target
526supports it, it will emit globals to the section specified.</p>
527
Chris Lattner2cbdc452005-11-06 08:02:57 +0000528<p>An explicit alignment may be specified for a global. If not present, or if
529the alignment is set to zero, the alignment of the global is set by the target
530to whatever it feels convenient. If an explicit alignment is specified, the
531global is forced to have at least that much alignment. All alignments must be
532a power of 2.</p>
533
Chris Lattnerfa730212004-12-09 16:11:40 +0000534</div>
535
536
537<!-- ======================================================================= -->
538<div class="doc_subsection">
539 <a name="functionstructure">Functions</a>
540</div>
541
542<div class="doc_text">
543
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000544<p>LLVM function definitions consist of an optional <a href="#linkage">linkage
545type</a>, an optional <a href="#callingconv">calling convention</a>, a return
Chris Lattner88f6c462005-11-12 00:45:07 +0000546type, a function name, a (possibly empty) argument list, an optional section,
547an optional alignment, an opening curly brace,
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000548a list of basic blocks, and a closing curly brace. LLVM function declarations
549are defined with the "<tt>declare</tt>" keyword, an optional <a
Chris Lattner2cbdc452005-11-06 08:02:57 +0000550href="#callingconv">calling convention</a>, a return type, a function name,
551a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000552
553<p>A function definition contains a list of basic blocks, forming the CFG for
554the function. Each basic block may optionally start with a label (giving the
555basic block a symbol table entry), contains a list of instructions, and ends
556with a <a href="#terminators">terminator</a> instruction (such as a branch or
557function return).</p>
558
John Criswelle4c57cc2005-05-12 16:52:32 +0000559<p>The first basic block in a program is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000560executed on entrance to the function, and it is not allowed to have predecessor
561basic blocks (i.e. there can not be any branches to the entry block of a
562function). Because the block can have no predecessors, it also cannot have any
563<a href="#i_phi">PHI nodes</a>.</p>
564
565<p>LLVM functions are identified by their name and type signature. Hence, two
566functions with the same name but different parameter lists or return values are
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000567considered different functions, and LLVM will resolve references to each
Chris Lattnerfa730212004-12-09 16:11:40 +0000568appropriately.</p>
569
Chris Lattner88f6c462005-11-12 00:45:07 +0000570<p>LLVM allows an explicit section to be specified for functions. If the target
571supports it, it will emit functions to the section specified.</p>
572
Chris Lattner2cbdc452005-11-06 08:02:57 +0000573<p>An explicit alignment may be specified for a function. If not present, or if
574the alignment is set to zero, the alignment of the function is set by the target
575to whatever it feels convenient. If an explicit alignment is specified, the
576function is forced to have at least that much alignment. All alignments must be
577a power of 2.</p>
578
Chris Lattnerfa730212004-12-09 16:11:40 +0000579</div>
580
581
582
Chris Lattner00950542001-06-06 20:29:01 +0000583<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000584<div class="doc_section"> <a name="typesystem">Type System</a> </div>
585<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000586
Misha Brukman9d0919f2003-11-08 01:05:38 +0000587<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000588
Misha Brukman9d0919f2003-11-08 01:05:38 +0000589<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000590intermediate representation. Being typed enables a number of
591optimizations to be performed on the IR directly, without having to do
592extra analyses on the side before the transformation. A strong type
593system makes it easier to read the generated code and enables novel
594analyses and transformations that are not feasible to perform on normal
595three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000596
597</div>
598
Chris Lattner00950542001-06-06 20:29:01 +0000599<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000600<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000601<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000602<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000603system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000604
Reid Spencerd3f876c2004-11-01 08:19:36 +0000605<table class="layout">
606 <tr class="layout">
607 <td class="left">
608 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000609 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000610 <tr><th>Type</th><th>Description</th></tr>
611 <tr><td><tt>void</tt></td><td>No value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000612 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
613 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
614 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
615 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
616 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000617 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000618 </tbody>
619 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000620 </td>
621 <td class="right">
622 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000623 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000624 <tr><th>Type</th><th>Description</th></tr>
625 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000626 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
627 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
628 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
629 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
630 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000631 </tbody>
632 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000633 </td>
634 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000635</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000636</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000637
Chris Lattner00950542001-06-06 20:29:01 +0000638<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000639<div class="doc_subsubsection"> <a name="t_classifications">Type
640Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000641<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000642<p>These different primitive types fall into a few useful
643classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000644
645<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000646 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000647 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000648 <tr>
649 <td><a name="t_signed">signed</a></td>
650 <td><tt>sbyte, short, int, long, float, double</tt></td>
651 </tr>
652 <tr>
653 <td><a name="t_unsigned">unsigned</a></td>
654 <td><tt>ubyte, ushort, uint, ulong</tt></td>
655 </tr>
656 <tr>
657 <td><a name="t_integer">integer</a></td>
658 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
659 </tr>
660 <tr>
661 <td><a name="t_integral">integral</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000662 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
663 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000664 </tr>
665 <tr>
666 <td><a name="t_floating">floating point</a></td>
667 <td><tt>float, double</tt></td>
668 </tr>
669 <tr>
670 <td><a name="t_firstclass">first class</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000671 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
672 float, double, <a href="#t_pointer">pointer</a>,
673 <a href="#t_packed">packed</a></tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000674 </tr>
675 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000676</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000677
Chris Lattner261efe92003-11-25 01:02:51 +0000678<p>The <a href="#t_firstclass">first class</a> types are perhaps the
679most important. Values of these types are the only ones which can be
680produced by instructions, passed as arguments, or used as operands to
681instructions. This means that all structures and arrays must be
682manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000683</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000684
Chris Lattner00950542001-06-06 20:29:01 +0000685<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000686<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000687
Misha Brukman9d0919f2003-11-08 01:05:38 +0000688<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000689
Chris Lattner261efe92003-11-25 01:02:51 +0000690<p>The real power in LLVM comes from the derived types in the system.
691This is what allows a programmer to represent arrays, functions,
692pointers, and other useful types. Note that these derived types may be
693recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000694
Misha Brukman9d0919f2003-11-08 01:05:38 +0000695</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000696
Chris Lattner00950542001-06-06 20:29:01 +0000697<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000698<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000699
Misha Brukman9d0919f2003-11-08 01:05:38 +0000700<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000701
Chris Lattner00950542001-06-06 20:29:01 +0000702<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000703
Misha Brukman9d0919f2003-11-08 01:05:38 +0000704<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000705sequentially in memory. The array type requires a size (number of
706elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000707
Chris Lattner7faa8832002-04-14 06:13:44 +0000708<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000709
710<pre>
711 [&lt;# elements&gt; x &lt;elementtype&gt;]
712</pre>
713
John Criswelle4c57cc2005-05-12 16:52:32 +0000714<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000715be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000716
Chris Lattner7faa8832002-04-14 06:13:44 +0000717<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000718<table class="layout">
719 <tr class="layout">
720 <td class="left">
721 <tt>[40 x int ]</tt><br/>
722 <tt>[41 x int ]</tt><br/>
723 <tt>[40 x uint]</tt><br/>
724 </td>
725 <td class="left">
726 Array of 40 integer values.<br/>
727 Array of 41 integer values.<br/>
728 Array of 40 unsigned integer values.<br/>
729 </td>
730 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000731</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000732<p>Here are some examples of multidimensional arrays:</p>
733<table class="layout">
734 <tr class="layout">
735 <td class="left">
736 <tt>[3 x [4 x int]]</tt><br/>
737 <tt>[12 x [10 x float]]</tt><br/>
738 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
739 </td>
740 <td class="left">
John Criswellc1f786c2005-05-13 22:25:59 +0000741 3x4 array of integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000742 12x10 array of single precision floating point values.<br/>
743 2x3x4 array of unsigned integer values.<br/>
744 </td>
745 </tr>
746</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000747
John Criswell0ec250c2005-10-24 16:17:18 +0000748<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
749length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000750LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
751As a special case, however, zero length arrays are recognized to be variable
752length. This allows implementation of 'pascal style arrays' with the LLVM
753type "{ int, [0 x float]}", for example.</p>
754
Misha Brukman9d0919f2003-11-08 01:05:38 +0000755</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000756
Chris Lattner00950542001-06-06 20:29:01 +0000757<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000758<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000759<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000760<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000761<p>The function type can be thought of as a function signature. It
762consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000763Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000764(which are structures of pointers to functions), for indirect function
765calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000766<p>
767The return type of a function type cannot be an aggregate type.
768</p>
Chris Lattner00950542001-06-06 20:29:01 +0000769<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000770<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000771<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +0000772specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000773which indicates that the function takes a variable number of arguments.
774Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000775 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000776<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000777<table class="layout">
778 <tr class="layout">
779 <td class="left">
780 <tt>int (int)</tt> <br/>
781 <tt>float (int, int *) *</tt><br/>
782 <tt>int (sbyte *, ...)</tt><br/>
783 </td>
784 <td class="left">
785 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
786 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukmanc24b7582004-08-12 20:16:08 +0000787 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
Reid Spencerd3f876c2004-11-01 08:19:36 +0000788 returning <tt>float</tt>.<br/>
789 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
790 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
791 the signature for <tt>printf</tt> in LLVM.<br/>
792 </td>
793 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000794</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000795
Misha Brukman9d0919f2003-11-08 01:05:38 +0000796</div>
Chris Lattner00950542001-06-06 20:29:01 +0000797<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000798<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000799<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000800<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000801<p>The structure type is used to represent a collection of data members
802together in memory. The packing of the field types is defined to match
803the ABI of the underlying processor. The elements of a structure may
804be any type that has a size.</p>
805<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
806and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
807field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
808instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000809<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000810<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000811<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000812<table class="layout">
813 <tr class="layout">
814 <td class="left">
815 <tt>{ int, int, int }</tt><br/>
816 <tt>{ float, int (int) * }</tt><br/>
817 </td>
818 <td class="left">
819 a triple of three <tt>int</tt> values<br/>
820 A pair, where the first element is a <tt>float</tt> and the second element
821 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
822 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
823 </td>
824 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000825</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000826</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000827
Chris Lattner00950542001-06-06 20:29:01 +0000828<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000829<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000830<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000831<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000832<p>As in many languages, the pointer type represents a pointer or
833reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000834<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000835<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000836<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000837<table class="layout">
838 <tr class="layout">
839 <td class="left">
840 <tt>[4x int]*</tt><br/>
841 <tt>int (int *) *</tt><br/>
842 </td>
843 <td class="left">
844 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
845 four <tt>int</tt> values<br/>
846 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattnera977c482005-02-19 02:22:14 +0000847 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerd3f876c2004-11-01 08:19:36 +0000848 <tt>int</tt>.<br/>
849 </td>
850 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000851</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000852</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000853
Chris Lattnera58561b2004-08-12 19:12:28 +0000854<!-- _______________________________________________________________________ -->
855<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000856<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000857
Chris Lattnera58561b2004-08-12 19:12:28 +0000858<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000859
Chris Lattnera58561b2004-08-12 19:12:28 +0000860<p>A packed type is a simple derived type that represents a vector
861of elements. Packed types are used when multiple primitive data
862are operated in parallel using a single instruction (SIMD).
863A packed type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000864elements) and an underlying primitive data type. Vectors must have a power
865of two length (1, 2, 4, 8, 16 ...). Packed types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000866considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000867
Chris Lattnera58561b2004-08-12 19:12:28 +0000868<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000869
870<pre>
871 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
872</pre>
873
John Criswellc1f786c2005-05-13 22:25:59 +0000874<p>The number of elements is a constant integer value; elementtype may
Chris Lattnera58561b2004-08-12 19:12:28 +0000875be any integral or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000876
Chris Lattnera58561b2004-08-12 19:12:28 +0000877<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000878
Reid Spencerd3f876c2004-11-01 08:19:36 +0000879<table class="layout">
880 <tr class="layout">
881 <td class="left">
882 <tt>&lt;4 x int&gt;</tt><br/>
883 <tt>&lt;8 x float&gt;</tt><br/>
884 <tt>&lt;2 x uint&gt;</tt><br/>
885 </td>
886 <td class="left">
887 Packed vector of 4 integer values.<br/>
888 Packed vector of 8 floating-point values.<br/>
889 Packed vector of 2 unsigned integer values.<br/>
890 </td>
891 </tr>
892</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000893</div>
894
Chris Lattner69c11bb2005-04-25 17:34:15 +0000895<!-- _______________________________________________________________________ -->
896<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
897<div class="doc_text">
898
899<h5>Overview:</h5>
900
901<p>Opaque types are used to represent unknown types in the system. This
902corresponds (for example) to the C notion of a foward declared structure type.
903In LLVM, opaque types can eventually be resolved to any type (not just a
904structure type).</p>
905
906<h5>Syntax:</h5>
907
908<pre>
909 opaque
910</pre>
911
912<h5>Examples:</h5>
913
914<table class="layout">
915 <tr class="layout">
916 <td class="left">
917 <tt>opaque</tt>
918 </td>
919 <td class="left">
920 An opaque type.<br/>
921 </td>
922 </tr>
923</table>
924</div>
925
926
Chris Lattnerc3f59762004-12-09 17:30:23 +0000927<!-- *********************************************************************** -->
928<div class="doc_section"> <a name="constants">Constants</a> </div>
929<!-- *********************************************************************** -->
930
931<div class="doc_text">
932
933<p>LLVM has several different basic types of constants. This section describes
934them all and their syntax.</p>
935
936</div>
937
938<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000939<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000940
941<div class="doc_text">
942
943<dl>
944 <dt><b>Boolean constants</b></dt>
945
946 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
947 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
948 </dd>
949
950 <dt><b>Integer constants</b></dt>
951
Reid Spencercc16dc32004-12-09 18:02:53 +0000952 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000953 href="#t_integer">integer</a> type. Negative numbers may be used with signed
954 integer types.
955 </dd>
956
957 <dt><b>Floating point constants</b></dt>
958
959 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
960 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000961 notation (see below). Floating point constants must have a <a
962 href="#t_floating">floating point</a> type. </dd>
963
964 <dt><b>Null pointer constants</b></dt>
965
John Criswell9e2485c2004-12-10 15:51:16 +0000966 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +0000967 and must be of <a href="#t_pointer">pointer type</a>.</dd>
968
969</dl>
970
John Criswell9e2485c2004-12-10 15:51:16 +0000971<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +0000972of floating point constants. For example, the form '<tt>double
9730x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
9744.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +0000975(and the only time that they are generated by the disassembler) is when a
976floating point constant must be emitted but it cannot be represented as a
977decimal floating point number. For example, NaN's, infinities, and other
978special values are represented in their IEEE hexadecimal format so that
979assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000980
981</div>
982
983<!-- ======================================================================= -->
984<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
985</div>
986
987<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000988<p>Aggregate constants arise from aggregation of simple constants
989and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000990
991<dl>
992 <dt><b>Structure constants</b></dt>
993
994 <dd>Structure constants are represented with notation similar to structure
995 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000996 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
997 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
998 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +0000999 types of elements must match those specified by the type.
1000 </dd>
1001
1002 <dt><b>Array constants</b></dt>
1003
1004 <dd>Array constants are represented with notation similar to array type
1005 definitions (a comma separated list of elements, surrounded by square brackets
John Criswell9e2485c2004-12-10 15:51:16 +00001006 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +00001007 constants must have <a href="#t_array">array type</a>, and the number and
1008 types of elements must match those specified by the type.
1009 </dd>
1010
1011 <dt><b>Packed constants</b></dt>
1012
1013 <dd>Packed constants are represented with notation similar to packed type
1014 definitions (a comma separated list of elements, surrounded by
John Criswell9e2485c2004-12-10 15:51:16 +00001015 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; int 42,
Chris Lattnerc3f59762004-12-09 17:30:23 +00001016 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
1017 href="#t_packed">packed type</a>, and the number and types of elements must
1018 match those specified by the type.
1019 </dd>
1020
1021 <dt><b>Zero initialization</b></dt>
1022
1023 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1024 value to zero of <em>any</em> type, including scalar and aggregate types.
1025 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +00001026 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +00001027 initializers.
1028 </dd>
1029</dl>
1030
1031</div>
1032
1033<!-- ======================================================================= -->
1034<div class="doc_subsection">
1035 <a name="globalconstants">Global Variable and Function Addresses</a>
1036</div>
1037
1038<div class="doc_text">
1039
1040<p>The addresses of <a href="#globalvars">global variables</a> and <a
1041href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +00001042constants. These constants are explicitly referenced when the <a
1043href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +00001044href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1045file:</p>
1046
1047<pre>
1048 %X = global int 17
1049 %Y = global int 42
1050 %Z = global [2 x int*] [ int* %X, int* %Y ]
1051</pre>
1052
1053</div>
1054
1055<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +00001056<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001057<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +00001058 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +00001059 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +00001060 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001061
Reid Spencer2dc45b82004-12-09 18:13:12 +00001062 <p>Undefined values indicate to the compiler that the program is well defined
1063 no matter what value is used, giving the compiler more freedom to optimize.
1064 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001065</div>
1066
1067<!-- ======================================================================= -->
1068<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1069</div>
1070
1071<div class="doc_text">
1072
1073<p>Constant expressions are used to allow expressions involving other constants
1074to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +00001075href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +00001076that does not have side effects (e.g. load and call are not supported). The
1077following is the syntax for constant expressions:</p>
1078
1079<dl>
1080 <dt><b><tt>cast ( CST to TYPE )</tt></b></dt>
1081
1082 <dd>Cast a constant to another type.</dd>
1083
1084 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1085
1086 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1087 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1088 instruction, the index list may have zero or more indexes, which are required
1089 to make sense for the type of "CSTPTR".</dd>
1090
Robert Bocchino9fbe1452006-01-10 19:31:34 +00001091 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1092
1093 <dd>Perform the <a href="#i_select">select operation</a> on
1094 constants.
1095
1096 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1097
1098 <dd>Perform the <a href="#i_extractelement">extractelement
1099 operation</a> on constants.
1100
Chris Lattnerc3f59762004-12-09 17:30:23 +00001101 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1102
Reid Spencer2dc45b82004-12-09 18:13:12 +00001103 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1104 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +00001105 binary</a> operations. The constraints on operands are the same as those for
1106 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +00001107 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001108</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001109</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00001110
Chris Lattner00950542001-06-06 20:29:01 +00001111<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00001112<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1113<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00001114
Misha Brukman9d0919f2003-11-08 01:05:38 +00001115<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001116
Chris Lattner261efe92003-11-25 01:02:51 +00001117<p>The LLVM instruction set consists of several different
1118classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +00001119instructions</a>, <a href="#binaryops">binary instructions</a>,
1120<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +00001121 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1122instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001123
Misha Brukman9d0919f2003-11-08 01:05:38 +00001124</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001125
Chris Lattner00950542001-06-06 20:29:01 +00001126<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001127<div class="doc_subsection"> <a name="terminators">Terminator
1128Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001129
Misha Brukman9d0919f2003-11-08 01:05:38 +00001130<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001131
Chris Lattner261efe92003-11-25 01:02:51 +00001132<p>As mentioned <a href="#functionstructure">previously</a>, every
1133basic block in a program ends with a "Terminator" instruction, which
1134indicates which block should be executed after the current block is
1135finished. These terminator instructions typically yield a '<tt>void</tt>'
1136value: they produce control flow, not values (the one exception being
1137the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +00001138<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +00001139 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1140instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +00001141the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1142 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1143 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001144
Misha Brukman9d0919f2003-11-08 01:05:38 +00001145</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001146
Chris Lattner00950542001-06-06 20:29:01 +00001147<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001148<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1149Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001150<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001151<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001152<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 +00001153 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001154</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001155<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001156<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +00001157value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +00001158<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +00001159returns a value and then causes control flow, and one that just causes
1160control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001161<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001162<p>The '<tt>ret</tt>' instruction may return any '<a
1163 href="#t_firstclass">first class</a>' type. Notice that a function is
1164not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1165instruction inside of the function that returns a value that does not
1166match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001167<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001168<p>When the '<tt>ret</tt>' instruction is executed, control flow
1169returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +00001170 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +00001171the instruction after the call. If the caller was an "<a
1172 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +00001173at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +00001174returns a value, that value shall set the call or invoke instruction's
1175return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001176<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001177<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001178 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001179</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001180</div>
Chris Lattner00950542001-06-06 20:29:01 +00001181<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001182<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001183<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001184<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001185<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 +00001186</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001187<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001188<p>The '<tt>br</tt>' instruction is used to cause control flow to
1189transfer to a different basic block in the current function. There are
1190two forms of this instruction, corresponding to a conditional branch
1191and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001192<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001193<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1194single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1195unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1196value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001197<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001198<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1199argument is evaluated. If the value is <tt>true</tt>, control flows
1200to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1201control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001202<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001203<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
1204 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 +00001205</div>
Chris Lattner00950542001-06-06 20:29:01 +00001206<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001207<div class="doc_subsubsection">
1208 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1209</div>
1210
Misha Brukman9d0919f2003-11-08 01:05:38 +00001211<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001212<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001213
1214<pre>
1215 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1216</pre>
1217
Chris Lattner00950542001-06-06 20:29:01 +00001218<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001219
1220<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1221several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +00001222instruction, allowing a branch to occur to one of many possible
1223destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001224
1225
Chris Lattner00950542001-06-06 20:29:01 +00001226<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001227
1228<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1229comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1230an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1231table is not allowed to contain duplicate constant entries.</p>
1232
Chris Lattner00950542001-06-06 20:29:01 +00001233<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001234
Chris Lattner261efe92003-11-25 01:02:51 +00001235<p>The <tt>switch</tt> instruction specifies a table of values and
1236destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +00001237table is searched for the given value. If the value is found, control flow is
1238transfered to the corresponding destination; otherwise, control flow is
1239transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001240
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001241<h5>Implementation:</h5>
1242
1243<p>Depending on properties of the target machine and the particular
1244<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +00001245ways. For example, it could be generated as a series of chained conditional
1246branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001247
1248<h5>Example:</h5>
1249
1250<pre>
1251 <i>; Emulate a conditional br instruction</i>
1252 %Val = <a href="#i_cast">cast</a> bool %value to int
1253 switch int %Val, label %truedest [int 0, label %falsedest ]
1254
1255 <i>; Emulate an unconditional br instruction</i>
1256 switch uint 0, label %dest [ ]
1257
1258 <i>; Implement a jump table:</i>
1259 switch uint %val, label %otherwise [ uint 0, label %onzero
1260 uint 1, label %onone
1261 uint 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00001262</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001263</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001264
Chris Lattner00950542001-06-06 20:29:01 +00001265<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001266<div class="doc_subsubsection">
1267 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1268</div>
1269
Misha Brukman9d0919f2003-11-08 01:05:38 +00001270<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001271
Chris Lattner00950542001-06-06 20:29:01 +00001272<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001273
1274<pre>
1275 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)
1276 to label &lt;normal label&gt; except label &lt;exception label&gt;
1277</pre>
1278
Chris Lattner6536cfe2002-05-06 22:08:29 +00001279<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001280
1281<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1282function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +00001283'<tt>normal</tt>' label or the
1284'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001285"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1286"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +00001287href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1288continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001289
Chris Lattner00950542001-06-06 20:29:01 +00001290<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001291
Misha Brukman9d0919f2003-11-08 01:05:38 +00001292<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001293
Chris Lattner00950542001-06-06 20:29:01 +00001294<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001295 <li>
John Criswellc1f786c2005-05-13 22:25:59 +00001296 The optional "cconv" marker indicates which <a href="callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001297 convention</a> the call should use. If none is specified, the call defaults
1298 to using C calling conventions.
1299 </li>
1300 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1301 function value being invoked. In most cases, this is a direct function
1302 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1303 an arbitrary pointer to function value.
1304 </li>
1305
1306 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1307 function to be invoked. </li>
1308
1309 <li>'<tt>function args</tt>': argument list whose types match the function
1310 signature argument types. If the function signature indicates the function
1311 accepts a variable number of arguments, the extra arguments can be
1312 specified. </li>
1313
1314 <li>'<tt>normal label</tt>': the label reached when the called function
1315 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1316
1317 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1318 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1319
Chris Lattner00950542001-06-06 20:29:01 +00001320</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001321
Chris Lattner00950542001-06-06 20:29:01 +00001322<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001323
Misha Brukman9d0919f2003-11-08 01:05:38 +00001324<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001325href="#i_call">call</a></tt>' instruction in most regards. The primary
1326difference is that it establishes an association with a label, which is used by
1327the runtime library to unwind the stack.</p>
1328
1329<p>This instruction is used in languages with destructors to ensure that proper
1330cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1331exception. Additionally, this is important for implementation of
1332'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1333
Chris Lattner00950542001-06-06 20:29:01 +00001334<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001335<pre>
1336 %retval = invoke int %Test(int 15) to label %Continue
1337 except label %TestCleanup <i>; {int}:retval set</i>
1338 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
1339 except label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00001340</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001341</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001342
1343
Chris Lattner27f71f22003-09-03 00:41:47 +00001344<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00001345
Chris Lattner261efe92003-11-25 01:02:51 +00001346<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1347Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +00001348
Misha Brukman9d0919f2003-11-08 01:05:38 +00001349<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +00001350
Chris Lattner27f71f22003-09-03 00:41:47 +00001351<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001352<pre>
1353 unwind
1354</pre>
1355
Chris Lattner27f71f22003-09-03 00:41:47 +00001356<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001357
1358<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1359at the first callee in the dynamic call stack which used an <a
1360href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1361primarily used to implement exception handling.</p>
1362
Chris Lattner27f71f22003-09-03 00:41:47 +00001363<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001364
1365<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1366immediately halt. The dynamic call stack is then searched for the first <a
1367href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1368execution continues at the "exceptional" destination block specified by the
1369<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1370dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001371</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001372
1373<!-- _______________________________________________________________________ -->
1374
1375<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1376Instruction</a> </div>
1377
1378<div class="doc_text">
1379
1380<h5>Syntax:</h5>
1381<pre>
1382 unreachable
1383</pre>
1384
1385<h5>Overview:</h5>
1386
1387<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1388instruction is used to inform the optimizer that a particular portion of the
1389code is not reachable. This can be used to indicate that the code after a
1390no-return function cannot be reached, and other facts.</p>
1391
1392<h5>Semantics:</h5>
1393
1394<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1395</div>
1396
1397
1398
Chris Lattner00950542001-06-06 20:29:01 +00001399<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001400<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001401<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001402<p>Binary operators are used to do most of the computation in a
1403program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +00001404produce a single value. The operands might represent
Chris Lattnera58561b2004-08-12 19:12:28 +00001405multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1406The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +00001407necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001408<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001409</div>
Chris Lattner00950542001-06-06 20:29:01 +00001410<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001411<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1412Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001413<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001414<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001415<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 +00001416</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001417<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001418<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001419<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001420<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +00001421 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1422 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1423Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001424<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001425<p>The value produced is the integer or floating point sum of the two
1426operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001427<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001428<pre> &lt;result&gt; = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001429</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001430</div>
Chris Lattner00950542001-06-06 20:29:01 +00001431<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001432<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1433Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001434<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001435<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001436<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 +00001437</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001438<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001439<p>The '<tt>sub</tt>' instruction returns the difference of its two
1440operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001441<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1442instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001443<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001444<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001445 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001446values.
1447This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1448Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001449<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001450<p>The value produced is the integer or floating point difference of
1451the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001452<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001453<pre> &lt;result&gt; = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001454 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1455</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001456</div>
Chris Lattner00950542001-06-06 20:29:01 +00001457<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001458<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1459Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001460<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001461<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001462<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 +00001463</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001464<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001465<p>The '<tt>mul</tt>' instruction returns the product of its two
1466operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001467<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001468<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001469 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001470values.
1471This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1472Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001473<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001474<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +00001475two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001476<p>There is no signed vs unsigned multiplication. The appropriate
1477action is taken based on the type of the operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001478<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001479<pre> &lt;result&gt; = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001480</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001481</div>
Chris Lattner00950542001-06-06 20:29:01 +00001482<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001483<div class="doc_subsubsection"> <a name="i_div">'<tt>div</tt>'
1484Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001485<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001486<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001487<pre> &lt;result&gt; = div &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1488</pre>
1489<h5>Overview:</h5>
1490<p>The '<tt>div</tt>' instruction returns the quotient of its two
1491operands.</p>
1492<h5>Arguments:</h5>
1493<p>The two arguments to the '<tt>div</tt>' instruction must be either <a
1494 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001495values.
1496This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1497Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001498<h5>Semantics:</h5>
1499<p>The value produced is the integer or floating point quotient of the
1500two operands.</p>
1501<h5>Example:</h5>
1502<pre> &lt;result&gt; = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1503</pre>
1504</div>
1505<!-- _______________________________________________________________________ -->
1506<div class="doc_subsubsection"> <a name="i_rem">'<tt>rem</tt>'
1507Instruction</a> </div>
1508<div class="doc_text">
1509<h5>Syntax:</h5>
1510<pre> &lt;result&gt; = rem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1511</pre>
1512<h5>Overview:</h5>
1513<p>The '<tt>rem</tt>' instruction returns the remainder from the
1514division of its two operands.</p>
1515<h5>Arguments:</h5>
1516<p>The two arguments to the '<tt>rem</tt>' instruction must be either <a
1517 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001518values.
1519This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1520Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001521<h5>Semantics:</h5>
1522<p>This returns the <i>remainder</i> of a division (where the result
1523has the same sign as the divisor), not the <i>modulus</i> (where the
1524result has the same sign as the dividend) of a value. For more
John Criswell0ec250c2005-10-24 16:17:18 +00001525information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +00001526 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1527Math Forum</a>.</p>
1528<h5>Example:</h5>
1529<pre> &lt;result&gt; = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1530</pre>
1531</div>
1532<!-- _______________________________________________________________________ -->
1533<div class="doc_subsubsection"> <a name="i_setcc">'<tt>set<i>cc</i></tt>'
1534Instructions</a> </div>
1535<div class="doc_text">
1536<h5>Syntax:</h5>
1537<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 +00001538 &lt;result&gt; = setne &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1539 &lt;result&gt; = setlt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1540 &lt;result&gt; = setgt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1541 &lt;result&gt; = setle &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1542 &lt;result&gt; = setge &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1543</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001544<h5>Overview:</h5>
1545<p>The '<tt>set<i>cc</i></tt>' family of instructions returns a boolean
1546value based on a comparison of their two operands.</p>
1547<h5>Arguments:</h5>
1548<p>The two arguments to the '<tt>set<i>cc</i></tt>' instructions must
1549be of <a href="#t_firstclass">first class</a> type (it is not possible
1550to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1551or '<tt>void</tt>' values, etc...). Both arguments must have identical
1552types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001553<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001554<p>The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1555value if both operands are equal.<br>
1556The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1557value if both operands are unequal.<br>
1558The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1559value if the first operand is less than the second operand.<br>
1560The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1561value if the first operand is greater than the second operand.<br>
1562The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1563value if the first operand is less than or equal to the second operand.<br>
1564The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1565value if the first operand is greater than or equal to the second
1566operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001567<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001568<pre> &lt;result&gt; = seteq int 4, 5 <i>; yields {bool}:result = false</i>
Chris Lattner00950542001-06-06 20:29:01 +00001569 &lt;result&gt; = setne float 4, 5 <i>; yields {bool}:result = true</i>
1570 &lt;result&gt; = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1571 &lt;result&gt; = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1572 &lt;result&gt; = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1573 &lt;result&gt; = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1574</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001575</div>
Chris Lattner00950542001-06-06 20:29:01 +00001576<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001577<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1578Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001579<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001580<p>Bitwise binary operators are used to do various forms of
1581bit-twiddling in a program. They are generally very efficient
John Criswell9e2485c2004-12-10 15:51:16 +00001582instructions and can commonly be strength reduced from other
Chris Lattner261efe92003-11-25 01:02:51 +00001583instructions. They require two operands, execute an operation on them,
1584and produce a single value. The resulting value of the bitwise binary
1585operators is always the same type as its first operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001586</div>
Chris Lattner00950542001-06-06 20:29:01 +00001587<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001588<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1589Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001590<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001591<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001592<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 +00001593</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001594<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001595<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1596its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001597<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001598<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001599 href="#t_integral">integral</a> values. Both arguments must have
1600identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001601<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001602<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001603<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001604<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +00001605<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001606 <tbody>
1607 <tr>
1608 <td>In0</td>
1609 <td>In1</td>
1610 <td>Out</td>
1611 </tr>
1612 <tr>
1613 <td>0</td>
1614 <td>0</td>
1615 <td>0</td>
1616 </tr>
1617 <tr>
1618 <td>0</td>
1619 <td>1</td>
1620 <td>0</td>
1621 </tr>
1622 <tr>
1623 <td>1</td>
1624 <td>0</td>
1625 <td>0</td>
1626 </tr>
1627 <tr>
1628 <td>1</td>
1629 <td>1</td>
1630 <td>1</td>
1631 </tr>
1632 </tbody>
1633</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001634</div>
Chris Lattner00950542001-06-06 20:29:01 +00001635<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001636<pre> &lt;result&gt; = and int 4, %var <i>; yields {int}:result = 4 &amp; %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001637 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1638 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1639</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001640</div>
Chris Lattner00950542001-06-06 20:29:01 +00001641<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001642<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001643<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001644<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001645<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 +00001646</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001647<h5>Overview:</h5>
1648<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1649or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001650<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001651<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001652 href="#t_integral">integral</a> values. Both arguments must have
1653identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001654<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001655<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001656<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001657<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001658<table border="1" cellspacing="0" cellpadding="4">
1659 <tbody>
1660 <tr>
1661 <td>In0</td>
1662 <td>In1</td>
1663 <td>Out</td>
1664 </tr>
1665 <tr>
1666 <td>0</td>
1667 <td>0</td>
1668 <td>0</td>
1669 </tr>
1670 <tr>
1671 <td>0</td>
1672 <td>1</td>
1673 <td>1</td>
1674 </tr>
1675 <tr>
1676 <td>1</td>
1677 <td>0</td>
1678 <td>1</td>
1679 </tr>
1680 <tr>
1681 <td>1</td>
1682 <td>1</td>
1683 <td>1</td>
1684 </tr>
1685 </tbody>
1686</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001687</div>
Chris Lattner00950542001-06-06 20:29:01 +00001688<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001689<pre> &lt;result&gt; = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001690 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1691 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1692</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001693</div>
Chris Lattner00950542001-06-06 20:29:01 +00001694<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001695<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1696Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001697<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001698<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001699<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 +00001700</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001701<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001702<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1703or of its two operands. The <tt>xor</tt> is used to implement the
1704"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001705<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001706<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001707 href="#t_integral">integral</a> values. Both arguments must have
1708identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001709<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001710<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001711<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001712<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001713<table border="1" cellspacing="0" cellpadding="4">
1714 <tbody>
1715 <tr>
1716 <td>In0</td>
1717 <td>In1</td>
1718 <td>Out</td>
1719 </tr>
1720 <tr>
1721 <td>0</td>
1722 <td>0</td>
1723 <td>0</td>
1724 </tr>
1725 <tr>
1726 <td>0</td>
1727 <td>1</td>
1728 <td>1</td>
1729 </tr>
1730 <tr>
1731 <td>1</td>
1732 <td>0</td>
1733 <td>1</td>
1734 </tr>
1735 <tr>
1736 <td>1</td>
1737 <td>1</td>
1738 <td>0</td>
1739 </tr>
1740 </tbody>
1741</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001742</div>
Chris Lattner261efe92003-11-25 01:02:51 +00001743<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +00001744<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001745<pre> &lt;result&gt; = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001746 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
1747 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner27f71f22003-09-03 00:41:47 +00001748 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00001749</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001750</div>
Chris Lattner00950542001-06-06 20:29:01 +00001751<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001752<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
1753Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001754<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001755<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001756<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 +00001757</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001758<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001759<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
1760the left a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001761<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001762<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001763 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1764type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001765<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001766<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001767<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001768<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 +00001769 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1770 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1771</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001772</div>
Chris Lattner00950542001-06-06 20:29:01 +00001773<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001774<div class="doc_subsubsection"> <a name="i_shr">'<tt>shr</tt>'
1775Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001776<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001777<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001778<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 +00001779</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001780<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001781<p>The '<tt>shr</tt>' instruction returns the first operand shifted to
1782the right a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001783<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001784<p>The first argument to the '<tt>shr</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001785 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1786type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001787<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001788<p>If the first argument is a <a href="#t_signed">signed</a> type, the
1789most significant bit is duplicated in the newly free'd bit positions.
1790If the first argument is unsigned, zero bits shall fill the empty
1791positions.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001792<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001793<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 +00001794 &lt;result&gt; = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
Chris Lattner00950542001-06-06 20:29:01 +00001795 &lt;result&gt; = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +00001796 &lt;result&gt; = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1797 &lt;result&gt; = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner00950542001-06-06 20:29:01 +00001798</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001799</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001800
Chris Lattner00950542001-06-06 20:29:01 +00001801<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001802<div class="doc_subsection">
1803 <a name="memoryops">Memory Access Operations</a>
1804</div>
1805
Misha Brukman9d0919f2003-11-08 01:05:38 +00001806<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001807
Chris Lattner261efe92003-11-25 01:02:51 +00001808<p>A key design point of an SSA-based representation is how it
1809represents memory. In LLVM, no memory locations are in SSA form, which
1810makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +00001811allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001812
Misha Brukman9d0919f2003-11-08 01:05:38 +00001813</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001814
Chris Lattner00950542001-06-06 20:29:01 +00001815<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001816<div class="doc_subsubsection">
1817 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
1818</div>
1819
Misha Brukman9d0919f2003-11-08 01:05:38 +00001820<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001821
Chris Lattner00950542001-06-06 20:29:01 +00001822<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001823
1824<pre>
1825 &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 +00001826</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001827
Chris Lattner00950542001-06-06 20:29:01 +00001828<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001829
Chris Lattner261efe92003-11-25 01:02:51 +00001830<p>The '<tt>malloc</tt>' instruction allocates memory from the system
1831heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001832
Chris Lattner00950542001-06-06 20:29:01 +00001833<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001834
1835<p>The '<tt>malloc</tt>' instruction allocates
1836<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +00001837bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00001838appropriate type to the program. If "NumElements" is specified, it is the
1839number of elements allocated. If an alignment is specified, the value result
1840of the allocation is guaranteed to be aligned to at least that boundary. If
1841not specified, or if zero, the target can choose to align the allocation on any
1842convenient boundary.</p>
1843
Misha Brukman9d0919f2003-11-08 01:05:38 +00001844<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001845
Chris Lattner00950542001-06-06 20:29:01 +00001846<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001847
Chris Lattner261efe92003-11-25 01:02:51 +00001848<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
1849a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001850
Chris Lattner2cbdc452005-11-06 08:02:57 +00001851<h5>Example:</h5>
1852
1853<pre>
1854 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
1855
1856 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001857 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1858 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001859 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
1860 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +00001861</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001862</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001863
Chris Lattner00950542001-06-06 20:29:01 +00001864<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001865<div class="doc_subsubsection">
1866 <a name="i_free">'<tt>free</tt>' Instruction</a>
1867</div>
1868
Misha Brukman9d0919f2003-11-08 01:05:38 +00001869<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001870
Chris Lattner00950542001-06-06 20:29:01 +00001871<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001872
1873<pre>
1874 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +00001875</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001876
Chris Lattner00950542001-06-06 20:29:01 +00001877<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001878
Chris Lattner261efe92003-11-25 01:02:51 +00001879<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +00001880memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001881
Chris Lattner00950542001-06-06 20:29:01 +00001882<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001883
Chris Lattner261efe92003-11-25 01:02:51 +00001884<p>'<tt>value</tt>' shall be a pointer value that points to a value
1885that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
1886instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001887
Chris Lattner00950542001-06-06 20:29:01 +00001888<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001889
John Criswell9e2485c2004-12-10 15:51:16 +00001890<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +00001891after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001892
Chris Lattner00950542001-06-06 20:29:01 +00001893<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001894
1895<pre>
1896 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
Chris Lattner00950542001-06-06 20:29:01 +00001897 free [4 x ubyte]* %array
1898</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001899</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001900
Chris Lattner00950542001-06-06 20:29:01 +00001901<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001902<div class="doc_subsubsection">
1903 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
1904</div>
1905
Misha Brukman9d0919f2003-11-08 01:05:38 +00001906<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001907
Chris Lattner00950542001-06-06 20:29:01 +00001908<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001909
1910<pre>
1911 &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 +00001912</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001913
Chris Lattner00950542001-06-06 20:29:01 +00001914<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001915
Chris Lattner261efe92003-11-25 01:02:51 +00001916<p>The '<tt>alloca</tt>' instruction allocates memory on the current
1917stack frame of the procedure that is live until the current function
1918returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001919
Chris Lattner00950542001-06-06 20:29:01 +00001920<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001921
John Criswell9e2485c2004-12-10 15:51:16 +00001922<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00001923bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00001924appropriate type to the program. If "NumElements" is specified, it is the
1925number of elements allocated. If an alignment is specified, the value result
1926of the allocation is guaranteed to be aligned to at least that boundary. If
1927not specified, or if zero, the target can choose to align the allocation on any
1928convenient boundary.</p>
1929
Misha Brukman9d0919f2003-11-08 01:05:38 +00001930<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001931
Chris Lattner00950542001-06-06 20:29:01 +00001932<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001933
John Criswellc1f786c2005-05-13 22:25:59 +00001934<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +00001935memory is automatically released when the function returns. The '<tt>alloca</tt>'
1936instruction is commonly used to represent automatic variables that must
1937have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +00001938 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001939instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001940
Chris Lattner00950542001-06-06 20:29:01 +00001941<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001942
1943<pre>
1944 %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001945 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001946 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
1947 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00001948</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001949</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001950
Chris Lattner00950542001-06-06 20:29:01 +00001951<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001952<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
1953Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001954<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00001955<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001956<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 +00001957<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001958<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001959<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001960<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +00001961address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +00001962 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +00001963marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +00001964the number or order of execution of this <tt>load</tt> with other
1965volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
1966instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001967<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001968<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001969<h5>Examples:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001970<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1971 <a
1972 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001973 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1974</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001975</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001976<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001977<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
1978Instruction</a> </div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001979<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001980<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 +00001981 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 +00001982</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001983<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001984<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001985<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001986<p>There are two arguments to the '<tt>store</tt>' instruction: a value
John Criswell0ec250c2005-10-24 16:17:18 +00001987to 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 +00001988operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +00001989operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +00001990optimizer is not allowed to modify the number or order of execution of
1991this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
1992 href="#i_store">store</a></tt> instructions.</p>
1993<h5>Semantics:</h5>
1994<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
1995at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00001996<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001997<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1998 <a
1999 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002000 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2001</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002002<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002003<div class="doc_subsubsection">
2004 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2005</div>
2006
Misha Brukman9d0919f2003-11-08 01:05:38 +00002007<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +00002008<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002009<pre>
2010 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2011</pre>
2012
Chris Lattner7faa8832002-04-14 06:13:44 +00002013<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002014
2015<p>
2016The '<tt>getelementptr</tt>' instruction is used to get the address of a
2017subelement of an aggregate data structure.</p>
2018
Chris Lattner7faa8832002-04-14 06:13:44 +00002019<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002020
2021<p>This instruction takes a list of integer constants that indicate what
2022elements of the aggregate object to index to. The actual types of the arguments
2023provided depend on the type of the first pointer argument. The
2024'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +00002025levels of a structure or to a specific index in an array. When indexing into a
2026structure, only <tt>uint</tt>
John Criswellc1f786c2005-05-13 22:25:59 +00002027integer constants are allowed. When indexing into an array or pointer,
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002028<tt>int</tt> and <tt>long</tt> indexes are allowed of any sign.</p>
2029
Chris Lattner261efe92003-11-25 01:02:51 +00002030<p>For example, let's consider a C code fragment and how it gets
2031compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002032
2033<pre>
2034 struct RT {
2035 char A;
2036 int B[10][20];
2037 char C;
2038 };
2039 struct ST {
2040 int X;
2041 double Y;
2042 struct RT Z;
2043 };
2044
2045 int *foo(struct ST *s) {
2046 return &amp;s[1].Z.B[5][13];
2047 }
2048</pre>
2049
Misha Brukman9d0919f2003-11-08 01:05:38 +00002050<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002051
2052<pre>
2053 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2054 %ST = type { int, double, %RT }
2055
Brian Gaeke7283e7c2004-07-02 21:08:14 +00002056 implementation
2057
2058 int* %foo(%ST* %s) {
2059 entry:
2060 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002061 ret int* %reg
2062 }
2063</pre>
2064
Chris Lattner7faa8832002-04-14 06:13:44 +00002065<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002066
2067<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +00002068on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Chris Lattnere53e5082004-06-03 22:57:15 +00002069and <a href="#t_array">array</a> types require <tt>uint</tt>, <tt>int</tt>,
2070<tt>ulong</tt>, or <tt>long</tt> values, and <a href="#t_struct">structure</a>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002071types require <tt>uint</tt> <b>constants</b>.</p>
2072
Misha Brukman9d0919f2003-11-08 01:05:38 +00002073<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002074type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2075}</tt>' type, a structure. The second index indexes into the third element of
2076the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2077sbyte }</tt>' type, another structure. The third index indexes into the second
2078element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2079array. The two dimensions of the array are subscripted into, yielding an
John Criswellfc6b8952005-05-16 16:17:45 +00002080'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002081to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2082
Chris Lattner261efe92003-11-25 01:02:51 +00002083<p>Note that it is perfectly legal to index partially through a
2084structure, returning a pointer to an inner element. Because of this,
2085the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002086
2087<pre>
Chris Lattnerd4f6b172005-03-07 22:13:59 +00002088 int* %foo(%ST* %s) {
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002089 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2090 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2091 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2092 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2093 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2094 ret int* %t5
2095 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00002096</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00002097
2098<p>Note that it is undefined to access an array out of bounds: array and
2099pointer indexes must always be within the defined bounds of the array type.
2100The one exception for this rules is zero length arrays. These arrays are
2101defined to be accessible as variable length arrays, which requires access
2102beyond the zero'th element.</p>
2103
Chris Lattner7faa8832002-04-14 06:13:44 +00002104<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +00002105
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002106<pre>
2107 <i>; yields [12 x ubyte]*:aptr</i>
2108 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2109</pre>
2110
2111</div>
Chris Lattner00950542001-06-06 20:29:01 +00002112<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00002113<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002114<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +00002115<p>The instructions in this category are the "miscellaneous"
Chris Lattner261efe92003-11-25 01:02:51 +00002116instructions, which defy better classification.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002117</div>
Chris Lattner00950542001-06-06 20:29:01 +00002118<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002119<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
2120Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002121<div class="doc_text">
Chris Lattner33ba0d92001-07-09 00:26:23 +00002122<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002123<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002124<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002125<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
2126the SSA graph representing the function.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002127<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002128<p>The type of the incoming values are specified with the first type
2129field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
2130as arguments, with one pair for each predecessor basic block of the
2131current block. Only values of <a href="#t_firstclass">first class</a>
2132type may be used as the value arguments to the PHI node. Only labels
2133may be used as the label arguments.</p>
2134<p>There must be no non-phi instructions between the start of a basic
2135block and the PHI instructions: i.e. PHI instructions must be first in
2136a basic block.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002137<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002138<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
2139value specified by the parameter, depending on which basic block we
2140came from in the last <a href="#terminators">terminator</a> instruction.</p>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002141<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002142<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 +00002143</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002144
Chris Lattner6536cfe2002-05-06 22:08:29 +00002145<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +00002146<div class="doc_subsubsection">
2147 <a name="i_cast">'<tt>cast .. to</tt>' Instruction</a>
2148</div>
2149
Misha Brukman9d0919f2003-11-08 01:05:38 +00002150<div class="doc_text">
Chris Lattnercc37aae2004-03-12 05:50:16 +00002151
Chris Lattner6536cfe2002-05-06 22:08:29 +00002152<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002153
2154<pre>
2155 &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 +00002156</pre>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002157
Chris Lattner6536cfe2002-05-06 22:08:29 +00002158<h5>Overview:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002159
2160<p>
2161The '<tt>cast</tt>' instruction is used as the primitive means to convert
2162integers to floating point, change data type sizes, and break type safety (by
2163casting pointers).
2164</p>
2165
2166
Chris Lattner6536cfe2002-05-06 22:08:29 +00002167<h5>Arguments:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002168
2169<p>
2170The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
2171class value, and a type to cast it to, which must also be a <a
2172href="#t_firstclass">first class</a> type.
2173</p>
2174
Chris Lattner6536cfe2002-05-06 22:08:29 +00002175<h5>Semantics:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002176
2177<p>
2178This instruction follows the C rules for explicit casts when determining how the
2179data being cast must change to fit in its new container.
2180</p>
2181
2182<p>
2183When casting to bool, any value that would be considered true in the context of
2184a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
2185all else are '<tt>false</tt>'.
2186</p>
2187
2188<p>
2189When extending an integral value from a type of one signness to another (for
2190example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
2191<b>source</b> value is signed, and zero-extended if the source value is
2192unsigned. <tt>bool</tt> values are always zero extended into either zero or
2193one.
2194</p>
2195
Chris Lattner33ba0d92001-07-09 00:26:23 +00002196<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002197
2198<pre>
2199 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
Chris Lattner7bae3952002-06-25 18:03:17 +00002200 %Y = cast int 123 to bool <i>; yields bool:true</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002201</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002202</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002203
2204<!-- _______________________________________________________________________ -->
2205<div class="doc_subsubsection">
2206 <a name="i_select">'<tt>select</tt>' Instruction</a>
2207</div>
2208
2209<div class="doc_text">
2210
2211<h5>Syntax:</h5>
2212
2213<pre>
2214 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
2215</pre>
2216
2217<h5>Overview:</h5>
2218
2219<p>
2220The '<tt>select</tt>' instruction is used to choose one value based on a
2221condition, without branching.
2222</p>
2223
2224
2225<h5>Arguments:</h5>
2226
2227<p>
2228The '<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.
2229</p>
2230
2231<h5>Semantics:</h5>
2232
2233<p>
2234If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +00002235value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +00002236</p>
2237
2238<h5>Example:</h5>
2239
2240<pre>
2241 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
2242</pre>
2243</div>
2244
2245
Robert Bocchino3a558662006-01-05 17:37:02 +00002246<!-- _______________________________________________________________________ -->
2247<div class="doc_subsubsection">
2248 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2249</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002250
Robert Bocchino3a558662006-01-05 17:37:02 +00002251<div class="doc_text">
2252
2253<h5>Syntax:</h5>
2254
2255<pre>
2256 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, uint &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2257</pre>
2258
2259<h5>Overview:</h5>
2260
2261<p>
2262The '<tt>extractelement</tt>' instruction extracts a single scalar
2263element from a vector at a specified index.
2264</p>
2265
2266
2267<h5>Arguments:</h5>
2268
2269<p>
2270The first operand of an '<tt>extractelement</tt>' instruction is a
2271value of <a href="#t_packed">packed</a> type. The second operand is
2272an index indicating the position from which to extract the element.
2273The index may be a variable.</p>
2274
2275<h5>Semantics:</h5>
2276
2277<p>
2278The result is a scalar of the same type as the element type of
2279<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2280<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2281results are undefined.
2282</p>
2283
2284<h5>Example:</h5>
2285
2286<pre>
2287 %result = extractelement &lt;4 x int&gt; %vec, uint 0 <i>; yields int</i>
2288</pre>
2289</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002290
2291
Chris Lattner33ba0d92001-07-09 00:26:23 +00002292<!-- _______________________________________________________________________ -->
Chris Lattner2bff5242005-05-06 05:47:36 +00002293<div class="doc_subsubsection">
2294 <a name="i_call">'<tt>call</tt>' Instruction</a>
2295</div>
2296
Misha Brukman9d0919f2003-11-08 01:05:38 +00002297<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +00002298
Chris Lattner00950542001-06-06 20:29:01 +00002299<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002300<pre>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002301 &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 +00002302</pre>
2303
Chris Lattner00950542001-06-06 20:29:01 +00002304<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002305
Misha Brukman9d0919f2003-11-08 01:05:38 +00002306<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002307
Chris Lattner00950542001-06-06 20:29:01 +00002308<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002309
Misha Brukman9d0919f2003-11-08 01:05:38 +00002310<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002311
Chris Lattner6536cfe2002-05-06 22:08:29 +00002312<ol>
Chris Lattner261efe92003-11-25 01:02:51 +00002313 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002314 <p>The optional "tail" marker indicates whether the callee function accesses
2315 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +00002316 function call is eligible for tail call optimization. Note that calls may
2317 be marked "tail" even if they do not occur before a <a
2318 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +00002319 </li>
2320 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002321 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
2322 convention</a> the call should use. If none is specified, the call defaults
2323 to using C calling conventions.
2324 </li>
2325 <li>
Chris Lattner2bff5242005-05-06 05:47:36 +00002326 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
2327 being invoked. The argument types must match the types implied by this
John Criswellfc6b8952005-05-16 16:17:45 +00002328 signature. This type can be omitted if the function is not varargs and
2329 if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002330 </li>
2331 <li>
2332 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
2333 be invoked. In most cases, this is a direct function invocation, but
2334 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +00002335 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002336 </li>
2337 <li>
2338 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +00002339 function signature argument types. All arguments must be of
2340 <a href="#t_firstclass">first class</a> type. If the function signature
2341 indicates the function accepts a variable number of arguments, the extra
2342 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002343 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002344</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00002345
Chris Lattner00950542001-06-06 20:29:01 +00002346<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002347
Chris Lattner261efe92003-11-25 01:02:51 +00002348<p>The '<tt>call</tt>' instruction is used to cause control flow to
2349transfer to a specified function, with its incoming arguments bound to
2350the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
2351instruction in the called function, control flow continues with the
2352instruction after the function call, and the return value of the
2353function is bound to the result argument. This is a simpler case of
2354the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002355
Chris Lattner00950542001-06-06 20:29:01 +00002356<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002357
2358<pre>
2359 %retval = call int %test(int %argc)
2360 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
2361 %X = tail call int %foo()
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002362 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
Chris Lattner2bff5242005-05-06 05:47:36 +00002363</pre>
2364
Misha Brukman9d0919f2003-11-08 01:05:38 +00002365</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002366
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002367<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +00002368<div class="doc_subsubsection">
Chris Lattnere19d7a72004-09-27 21:51:25 +00002369 <a name="i_vaarg">'<tt>vaarg</tt>' Instruction</a>
2370</div>
2371
Misha Brukman9d0919f2003-11-08 01:05:38 +00002372<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +00002373
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002374<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002375
2376<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002377 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00002378</pre>
2379
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002380<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002381
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002382<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +00002383the "variable argument" area of a function call. It is used to implement the
2384<tt>va_arg</tt> macro in C.</p>
2385
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002386<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002387
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002388<p>This instruction takes a <tt>va_list*</tt> value and the type of
2389the argument. It returns a value of the specified argument type and
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00002390increments the <tt>va_list</tt> to point to the next argument. Again, the
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002391actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002392
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002393<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002394
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002395<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
2396type from the specified <tt>va_list</tt> and causes the
2397<tt>va_list</tt> to point to the next argument. For more information,
2398see the variable argument handling <a href="#int_varargs">Intrinsic
2399Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002400
2401<p>It is legal for this instruction to be called in a function which does not
2402take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002403function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002404
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002405<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +00002406href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +00002407argument.</p>
2408
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002409<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002410
2411<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
2412
Misha Brukman9d0919f2003-11-08 01:05:38 +00002413</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002414
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002415<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00002416<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
2417<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002418
Misha Brukman9d0919f2003-11-08 01:05:38 +00002419<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +00002420
2421<p>LLVM supports the notion of an "intrinsic function". These functions have
John Criswellfc6b8952005-05-16 16:17:45 +00002422well known names and semantics and are required to follow certain
Chris Lattner33aec9e2004-02-12 17:01:32 +00002423restrictions. Overall, these instructions represent an extension mechanism for
2424the LLVM language that does not require changing all of the transformations in
2425LLVM to add to the language (or the bytecode reader/writer, the parser,
2426etc...).</p>
2427
John Criswellfc6b8952005-05-16 16:17:45 +00002428<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
2429prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
Chris Lattner33aec9e2004-02-12 17:01:32 +00002430this. Intrinsic functions must always be external functions: you cannot define
2431the body of intrinsic functions. Intrinsic functions may only be used in call
2432or invoke instructions: it is illegal to take the address of an intrinsic
2433function. Additionally, because intrinsic functions are part of the LLVM
2434language, it is required that they all be documented here if any are added.</p>
2435
2436
John Criswellfc6b8952005-05-16 16:17:45 +00002437<p>To learn how to add an intrinsic function, please see the <a
Chris Lattner590cff32005-05-11 03:35:57 +00002438href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +00002439</p>
2440
Misha Brukman9d0919f2003-11-08 01:05:38 +00002441</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002442
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002443<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002444<div class="doc_subsection">
2445 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
2446</div>
2447
Misha Brukman9d0919f2003-11-08 01:05:38 +00002448<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002449
Misha Brukman9d0919f2003-11-08 01:05:38 +00002450<p>Variable argument support is defined in LLVM with the <a
Chris Lattner261efe92003-11-25 01:02:51 +00002451 href="#i_vanext"><tt>vanext</tt></a> instruction and these three
2452intrinsic functions. These functions are related to the similarly
2453named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002454
Chris Lattner261efe92003-11-25 01:02:51 +00002455<p>All of these functions operate on arguments that use a
2456target-specific value type "<tt>va_list</tt>". The LLVM assembly
2457language reference manual does not define what this type is, so all
2458transformations should be prepared to handle intrinsics with any type
2459used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002460
Misha Brukman9d0919f2003-11-08 01:05:38 +00002461<p>This example shows how the <a href="#i_vanext"><tt>vanext</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +00002462instruction and the variable argument handling intrinsic functions are
2463used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002464
Chris Lattner33aec9e2004-02-12 17:01:32 +00002465<pre>
2466int %test(int %X, ...) {
2467 ; Initialize variable argument processing
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002468 %ap = alloca sbyte*
2469 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002470
2471 ; Read a single integer argument
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002472 %tmp = va_arg sbyte** %ap, int
Chris Lattner33aec9e2004-02-12 17:01:32 +00002473
2474 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002475 %aq = alloca sbyte*
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002476 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002477 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002478
2479 ; Stop processing of arguments.
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002480 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002481 ret int %tmp
2482}
2483</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002484</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002485
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002486<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002487<div class="doc_subsubsection">
2488 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
2489</div>
2490
2491
Misha Brukman9d0919f2003-11-08 01:05:38 +00002492<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002493<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002494<pre> declare void %llvm.va_start(&lt;va_list&gt;* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002495<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002496<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
2497<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
2498href="#i_va_arg">va_arg</a></tt>.</p>
2499
2500<h5>Arguments:</h5>
2501
2502<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
2503
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002504<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002505
2506<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
2507macro available in C. In a target-dependent way, it initializes the
2508<tt>va_list</tt> element the argument points to, so that the next call to
2509<tt>va_arg</tt> will produce the first variable argument passed to the function.
2510Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
2511last argument of the function, the compiler can figure that out.</p>
2512
Misha Brukman9d0919f2003-11-08 01:05:38 +00002513</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002514
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002515<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002516<div class="doc_subsubsection">
2517 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
2518</div>
2519
Misha Brukman9d0919f2003-11-08 01:05:38 +00002520<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002521<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002522<pre> declare void %llvm.va_end(&lt;va_list*&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002523<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002524<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
2525which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
2526or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002527<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002528<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002529<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002530<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002531macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
2532Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
2533 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
2534with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002535</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002536
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002537<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002538<div class="doc_subsubsection">
2539 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
2540</div>
2541
Misha Brukman9d0919f2003-11-08 01:05:38 +00002542<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002543
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002544<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002545
2546<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002547 declare void %llvm.va_copy(&lt;va_list&gt;* &lt;destarglist&gt;,
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002548 &lt;va_list&gt;* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00002549</pre>
2550
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002551<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002552
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002553<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
2554the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002555
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002556<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002557
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002558<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002559The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002560
Chris Lattnerd7923912004-05-23 21:06:01 +00002561
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002562<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002563
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002564<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
2565available in C. In a target-dependent way, it copies the source
2566<tt>va_list</tt> element into the destination list. This intrinsic is necessary
2567because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
Chris Lattnerd7923912004-05-23 21:06:01 +00002568arbitrarily complex and require memory allocation, for example.</p>
2569
Misha Brukman9d0919f2003-11-08 01:05:38 +00002570</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002571
Chris Lattner33aec9e2004-02-12 17:01:32 +00002572<!-- ======================================================================= -->
2573<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +00002574 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
2575</div>
2576
2577<div class="doc_text">
2578
2579<p>
2580LLVM support for <a href="GarbageCollection.html">Accurate Garbage
2581Collection</a> requires the implementation and generation of these intrinsics.
2582These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
2583stack</a>, as well as garbage collector implementations that require <a
2584href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
2585Front-ends for type-safe garbage collected languages should generate these
2586intrinsics to make use of the LLVM garbage collectors. For more details, see <a
2587href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
2588</p>
2589</div>
2590
2591<!-- _______________________________________________________________________ -->
2592<div class="doc_subsubsection">
2593 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
2594</div>
2595
2596<div class="doc_text">
2597
2598<h5>Syntax:</h5>
2599
2600<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002601 declare void %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00002602</pre>
2603
2604<h5>Overview:</h5>
2605
John Criswell9e2485c2004-12-10 15:51:16 +00002606<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +00002607the code generator, and allows some metadata to be associated with it.</p>
2608
2609<h5>Arguments:</h5>
2610
2611<p>The first argument specifies the address of a stack object that contains the
2612root pointer. The second pointer (which must be either a constant or a global
2613value address) contains the meta-data to be associated with the root.</p>
2614
2615<h5>Semantics:</h5>
2616
2617<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
2618location. At compile-time, the code generator generates information to allow
2619the runtime to find the pointer at GC safe points.
2620</p>
2621
2622</div>
2623
2624
2625<!-- _______________________________________________________________________ -->
2626<div class="doc_subsubsection">
2627 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
2628</div>
2629
2630<div class="doc_text">
2631
2632<h5>Syntax:</h5>
2633
2634<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002635 declare sbyte* %llvm.gcread(sbyte** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00002636</pre>
2637
2638<h5>Overview:</h5>
2639
2640<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
2641locations, allowing garbage collector implementations that require read
2642barriers.</p>
2643
2644<h5>Arguments:</h5>
2645
2646<p>The argument is the address to read from, which should be an address
2647allocated from the garbage collector.</p>
2648
2649<h5>Semantics:</h5>
2650
2651<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
2652instruction, but may be replaced with substantially more complex code by the
2653garbage collector runtime, as needed.</p>
2654
2655</div>
2656
2657
2658<!-- _______________________________________________________________________ -->
2659<div class="doc_subsubsection">
2660 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
2661</div>
2662
2663<div class="doc_text">
2664
2665<h5>Syntax:</h5>
2666
2667<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002668 declare void %llvm.gcwrite(sbyte* %P1, sbyte** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00002669</pre>
2670
2671<h5>Overview:</h5>
2672
2673<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
2674locations, allowing garbage collector implementations that require write
2675barriers (such as generational or reference counting collectors).</p>
2676
2677<h5>Arguments:</h5>
2678
2679<p>The first argument is the reference to store, and the second is the heap
2680location to store to.</p>
2681
2682<h5>Semantics:</h5>
2683
2684<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
2685instruction, but may be replaced with substantially more complex code by the
2686garbage collector runtime, as needed.</p>
2687
2688</div>
2689
2690
2691
2692<!-- ======================================================================= -->
2693<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +00002694 <a name="int_codegen">Code Generator Intrinsics</a>
2695</div>
2696
2697<div class="doc_text">
2698<p>
2699These intrinsics are provided by LLVM to expose special features that may only
2700be implemented with code generator support.
2701</p>
2702
2703</div>
2704
2705<!-- _______________________________________________________________________ -->
2706<div class="doc_subsubsection">
2707 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
2708</div>
2709
2710<div class="doc_text">
2711
2712<h5>Syntax:</h5>
2713<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00002714 declare sbyte *%llvm.returnaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00002715</pre>
2716
2717<h5>Overview:</h5>
2718
2719<p>
2720The '<tt>llvm.returnaddress</tt>' intrinsic returns a target-specific value
2721indicating the return address of the current function or one of its callers.
2722</p>
2723
2724<h5>Arguments:</h5>
2725
2726<p>
2727The argument to this intrinsic indicates which function to return the address
2728for. Zero indicates the calling function, one indicates its caller, etc. The
2729argument is <b>required</b> to be a constant integer value.
2730</p>
2731
2732<h5>Semantics:</h5>
2733
2734<p>
2735The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
2736the return address of the specified call frame, or zero if it cannot be
2737identified. The value returned by this intrinsic is likely to be incorrect or 0
2738for arguments other than zero, so it should only be used for debugging purposes.
2739</p>
2740
2741<p>
2742Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00002743aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00002744source-language caller.
2745</p>
2746</div>
2747
2748
2749<!-- _______________________________________________________________________ -->
2750<div class="doc_subsubsection">
2751 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
2752</div>
2753
2754<div class="doc_text">
2755
2756<h5>Syntax:</h5>
2757<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00002758 declare sbyte *%llvm.frameaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00002759</pre>
2760
2761<h5>Overview:</h5>
2762
2763<p>
2764The '<tt>llvm.frameaddress</tt>' intrinsic returns the target-specific frame
2765pointer value for the specified stack frame.
2766</p>
2767
2768<h5>Arguments:</h5>
2769
2770<p>
2771The argument to this intrinsic indicates which function to return the frame
2772pointer for. Zero indicates the calling function, one indicates its caller,
2773etc. The argument is <b>required</b> to be a constant integer value.
2774</p>
2775
2776<h5>Semantics:</h5>
2777
2778<p>
2779The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
2780the frame address of the specified call frame, or zero if it cannot be
2781identified. The value returned by this intrinsic is likely to be incorrect or 0
2782for arguments other than zero, so it should only be used for debugging purposes.
2783</p>
2784
2785<p>
2786Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00002787aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00002788source-language caller.
2789</p>
2790</div>
2791
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002792<!-- _______________________________________________________________________ -->
2793<div class="doc_subsubsection">
Chris Lattner57e1f392006-01-13 02:03:13 +00002794 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
2795</div>
2796
2797<div class="doc_text">
2798
2799<h5>Syntax:</h5>
2800<pre>
2801 declare sbyte *%llvm.stacksave()
2802</pre>
2803
2804<h5>Overview:</h5>
2805
2806<p>
2807The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
2808the function stack, for use with <a href="#i_stackrestore">
2809<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
2810features like scoped automatic variable sized arrays in C99.
2811</p>
2812
2813<h5>Semantics:</h5>
2814
2815<p>
2816This intrinsic returns a opaque pointer value that can be passed to <a
2817href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
2818<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
2819<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
2820state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
2821practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
2822that were allocated after the <tt>llvm.stacksave</tt> was executed.
2823</p>
2824
2825</div>
2826
2827<!-- _______________________________________________________________________ -->
2828<div class="doc_subsubsection">
2829 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
2830</div>
2831
2832<div class="doc_text">
2833
2834<h5>Syntax:</h5>
2835<pre>
2836 declare void %llvm.stackrestore(sbyte* %ptr)
2837</pre>
2838
2839<h5>Overview:</h5>
2840
2841<p>
2842The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
2843the function stack to the state it was in when the corresponding <a
2844href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
2845useful for implementing language features like scoped automatic variable sized
2846arrays in C99.
2847</p>
2848
2849<h5>Semantics:</h5>
2850
2851<p>
2852See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
2853</p>
2854
2855</div>
2856
2857
2858<!-- _______________________________________________________________________ -->
2859<div class="doc_subsubsection">
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002860 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
2861</div>
2862
2863<div class="doc_text">
2864
2865<h5>Syntax:</h5>
2866<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002867 declare void %llvm.prefetch(sbyte * &lt;address&gt;,
2868 uint &lt;rw&gt;, uint &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002869</pre>
2870
2871<h5>Overview:</h5>
2872
2873
2874<p>
2875The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +00002876a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
2877no
2878effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +00002879characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002880</p>
2881
2882<h5>Arguments:</h5>
2883
2884<p>
2885<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
2886determining if the fetch should be for a read (0) or write (1), and
2887<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +00002888locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002889<tt>locality</tt> arguments must be constant integers.
2890</p>
2891
2892<h5>Semantics:</h5>
2893
2894<p>
2895This intrinsic does not modify the behavior of the program. In particular,
2896prefetches cannot trap and do not produce a value. On targets that support this
2897intrinsic, the prefetch can provide hints to the processor cache for better
2898performance.
2899</p>
2900
2901</div>
2902
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002903<!-- _______________________________________________________________________ -->
2904<div class="doc_subsubsection">
2905 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
2906</div>
2907
2908<div class="doc_text">
2909
2910<h5>Syntax:</h5>
2911<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002912 declare void %llvm.pcmarker( uint &lt;id&gt; )
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002913</pre>
2914
2915<h5>Overview:</h5>
2916
2917
2918<p>
John Criswellfc6b8952005-05-16 16:17:45 +00002919The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
2920(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002921code to simulators and other tools. The method is target specific, but it is
2922expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00002923The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +00002924after optimizations. It is possible that the presence of a marker will inhibit
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002925optimizations. The intended use is to be inserted after optmizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +00002926correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00002927</p>
2928
2929<h5>Arguments:</h5>
2930
2931<p>
2932<tt>id</tt> is a numerical id identifying the marker.
2933</p>
2934
2935<h5>Semantics:</h5>
2936
2937<p>
2938This intrinsic does not modify the behavior of the program. Backends that do not
2939support this intrinisic may ignore it.
2940</p>
2941
2942</div>
2943
Andrew Lenharth51b8d542005-11-11 16:47:30 +00002944<!-- _______________________________________________________________________ -->
2945<div class="doc_subsubsection">
2946 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
2947</div>
2948
2949<div class="doc_text">
2950
2951<h5>Syntax:</h5>
2952<pre>
2953 declare ulong %llvm.readcyclecounter( )
2954</pre>
2955
2956<h5>Overview:</h5>
2957
2958
2959<p>
2960The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
2961counter register (or similar low latency, high accuracy clocks) on those targets
2962that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
2963As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
2964should only be used for small timings.
2965</p>
2966
2967<h5>Semantics:</h5>
2968
2969<p>
2970When directly supported, reading the cycle counter should not modify any memory.
2971Implementations are allowed to either return a application specific value or a
2972system wide value. On backends without support, this is lowered to a constant 0.
2973</p>
2974
2975</div>
2976
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00002977
John Criswell7123e272004-04-09 16:43:20 +00002978<!-- ======================================================================= -->
2979<div class="doc_subsection">
2980 <a name="int_os">Operating System Intrinsics</a>
2981</div>
2982
2983<div class="doc_text">
2984<p>
2985These intrinsics are provided by LLVM to support the implementation of
2986operating system level code.
2987</p>
2988
2989</div>
John Criswell183402a2004-04-12 15:02:16 +00002990
John Criswellcfd3bac2004-04-09 15:23:37 +00002991<!-- _______________________________________________________________________ -->
2992<div class="doc_subsubsection">
2993 <a name="i_readport">'<tt>llvm.readport</tt>' Intrinsic</a>
2994</div>
2995
2996<div class="doc_text">
2997
2998<h5>Syntax:</h5>
2999<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003000 declare &lt;integer type&gt; %llvm.readport (&lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +00003001</pre>
3002
3003<h5>Overview:</h5>
3004
3005<p>
John Criswell7123e272004-04-09 16:43:20 +00003006The '<tt>llvm.readport</tt>' intrinsic reads data from the specified hardware
3007I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +00003008</p>
3009
3010<h5>Arguments:</h5>
3011
3012<p>
John Criswell7123e272004-04-09 16:43:20 +00003013The argument to this intrinsic indicates the hardware I/O address from which
3014to read the data. The address is in the hardware I/O address namespace (as
3015opposed to being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +00003016</p>
3017
3018<h5>Semantics:</h5>
3019
3020<p>
John Criswell7123e272004-04-09 16:43:20 +00003021The '<tt>llvm.readport</tt>' intrinsic reads data from the hardware I/O port
3022specified by <i>address</i> and returns the value. The address and return
3023value must be integers, but the size is dependent upon the platform upon which
3024the program is code generated. For example, on x86, the address must be an
Misha Brukmancfa87bc2005-04-22 18:02:52 +00003025unsigned 16-bit value, and the return value must be 8, 16, or 32 bits.
John Criswellcfd3bac2004-04-09 15:23:37 +00003026</p>
3027
3028</div>
3029
3030<!-- _______________________________________________________________________ -->
3031<div class="doc_subsubsection">
3032 <a name="i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a>
3033</div>
3034
3035<div class="doc_text">
3036
3037<h5>Syntax:</h5>
3038<pre>
Chris Lattnerc3f59762004-12-09 17:30:23 +00003039 call void (&lt;integer type&gt;, &lt;integer type&gt;)*
3040 %llvm.writeport (&lt;integer type&gt; &lt;value&gt;,
3041 &lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +00003042</pre>
3043
3044<h5>Overview:</h5>
3045
3046<p>
John Criswell7123e272004-04-09 16:43:20 +00003047The '<tt>llvm.writeport</tt>' intrinsic writes data to the specified hardware
3048I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +00003049</p>
3050
3051<h5>Arguments:</h5>
3052
3053<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003054The first argument is the value to write to the I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +00003055</p>
3056
3057<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003058The second argument indicates the hardware I/O address to which data should be
3059written. The address is in the hardware I/O address namespace (as opposed to
3060being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +00003061</p>
3062
3063<h5>Semantics:</h5>
3064
3065<p>
3066The '<tt>llvm.writeport</tt>' intrinsic writes <i>value</i> to the I/O port
3067specified by <i>address</i>. The address and value must be integers, but the
3068size is dependent upon the platform upon which the program is code generated.
Misha Brukmancfa87bc2005-04-22 18:02:52 +00003069For example, on x86, the address must be an unsigned 16-bit value, and the
John Criswell7123e272004-04-09 16:43:20 +00003070value written must be 8, 16, or 32 bits in length.
John Criswellcfd3bac2004-04-09 15:23:37 +00003071</p>
3072
3073</div>
Chris Lattner10610642004-02-14 04:08:35 +00003074
John Criswell183402a2004-04-12 15:02:16 +00003075<!-- _______________________________________________________________________ -->
3076<div class="doc_subsubsection">
3077 <a name="i_readio">'<tt>llvm.readio</tt>' Intrinsic</a>
3078</div>
3079
3080<div class="doc_text">
3081
3082<h5>Syntax:</h5>
3083<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003084 declare &lt;result&gt; %llvm.readio (&lt;ty&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +00003085</pre>
3086
3087<h5>Overview:</h5>
3088
3089<p>
3090The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
3091address.
3092</p>
3093
3094<h5>Arguments:</h5>
3095
3096<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003097The argument to this intrinsic is a pointer indicating the memory address from
3098which to read the data. The data must be a
3099<a href="#t_firstclass">first class</a> type.
John Criswell183402a2004-04-12 15:02:16 +00003100</p>
3101
3102<h5>Semantics:</h5>
3103
3104<p>
3105The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
John Criswell96db6fc2004-04-12 16:33:19 +00003106location specified by <i>pointer</i> and returns the value. The argument must
3107be a pointer, and the return value must be a
3108<a href="#t_firstclass">first class</a> type. However, certain architectures
Misha Brukmancfa87bc2005-04-22 18:02:52 +00003109may not support I/O on all first class types. For example, 32-bit processors
John Criswell96db6fc2004-04-12 16:33:19 +00003110may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +00003111</p>
3112
3113<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003114This intrinsic enforces an in-order memory model for llvm.readio and
3115llvm.writeio calls on machines that use dynamic scheduling. Dynamically
3116scheduled processors may execute loads and stores out of order, re-ordering at
3117run time accesses to memory mapped I/O registers. Using these intrinsics
3118ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +00003119</p>
3120
3121</div>
3122
3123<!-- _______________________________________________________________________ -->
3124<div class="doc_subsubsection">
3125 <a name="i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a>
3126</div>
3127
3128<div class="doc_text">
3129
3130<h5>Syntax:</h5>
3131<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003132 declare void %llvm.writeio (&lt;ty1&gt; &lt;value&gt;, &lt;ty2&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +00003133</pre>
3134
3135<h5>Overview:</h5>
3136
3137<p>
3138The '<tt>llvm.writeio</tt>' intrinsic writes data to the specified memory
3139mapped I/O address.
3140</p>
3141
3142<h5>Arguments:</h5>
3143
3144<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003145The first argument is the value to write to the memory mapped I/O location.
3146The second argument is a pointer indicating the memory address to which the
3147data should be written.
John Criswell183402a2004-04-12 15:02:16 +00003148</p>
3149
3150<h5>Semantics:</h5>
3151
3152<p>
3153The '<tt>llvm.writeio</tt>' intrinsic writes <i>value</i> to the memory mapped
John Criswell96db6fc2004-04-12 16:33:19 +00003154I/O address specified by <i>pointer</i>. The value must be a
3155<a href="#t_firstclass">first class</a> type. However, certain architectures
Misha Brukmancfa87bc2005-04-22 18:02:52 +00003156may not support I/O on all first class types. For example, 32-bit processors
John Criswell96db6fc2004-04-12 16:33:19 +00003157may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +00003158</p>
3159
3160<p>
John Criswell96db6fc2004-04-12 16:33:19 +00003161This intrinsic enforces an in-order memory model for llvm.readio and
3162llvm.writeio calls on machines that use dynamic scheduling. Dynamically
3163scheduled processors may execute loads and stores out of order, re-ordering at
3164run time accesses to memory mapped I/O registers. Using these intrinsics
3165ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +00003166</p>
3167
3168</div>
3169
Chris Lattner10610642004-02-14 04:08:35 +00003170<!-- ======================================================================= -->
3171<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +00003172 <a name="int_libc">Standard C Library Intrinsics</a>
3173</div>
3174
3175<div class="doc_text">
3176<p>
Chris Lattner10610642004-02-14 04:08:35 +00003177LLVM provides intrinsics for a few important standard C library functions.
3178These intrinsics allow source-language front-ends to pass information about the
3179alignment of the pointer arguments to the code generator, providing opportunity
3180for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003181</p>
3182
3183</div>
3184
3185<!-- _______________________________________________________________________ -->
3186<div class="doc_subsubsection">
3187 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
3188</div>
3189
3190<div class="doc_text">
3191
3192<h5>Syntax:</h5>
3193<pre>
Reid Spencerd4622352005-04-26 20:41:16 +00003194 declare void %llvm.memcpy(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3195 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003196</pre>
3197
3198<h5>Overview:</h5>
3199
3200<p>
3201The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
3202location to the destination location.
3203</p>
3204
3205<p>
3206Note that, unlike the standard libc function, the <tt>llvm.memcpy</tt> intrinsic
3207does not return a value, and takes an extra alignment argument.
3208</p>
3209
3210<h5>Arguments:</h5>
3211
3212<p>
3213The first argument is a pointer to the destination, the second is a pointer to
3214the source. The third argument is an (arbitrarily sized) integer argument
3215specifying the number of bytes to copy, and the fourth argument is the alignment
3216of the source and destination locations.
3217</p>
3218
Chris Lattner3301ced2004-02-12 21:18:15 +00003219<p>
3220If the call to this intrinisic has an alignment value that is not 0 or 1, then
3221the caller guarantees that the size of the copy is a multiple of the alignment
3222and that both the source and destination pointers are aligned to that boundary.
3223</p>
3224
Chris Lattner33aec9e2004-02-12 17:01:32 +00003225<h5>Semantics:</h5>
3226
3227<p>
3228The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
3229location to the destination location, which are not allowed to overlap. It
3230copies "len" bytes of memory over. If the argument is known to be aligned to
3231some boundary, this can be specified as the fourth argument, otherwise it should
3232be set to 0 or 1.
3233</p>
3234</div>
3235
3236
Chris Lattner0eb51b42004-02-12 18:10:10 +00003237<!-- _______________________________________________________________________ -->
3238<div class="doc_subsubsection">
3239 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
3240</div>
3241
3242<div class="doc_text">
3243
3244<h5>Syntax:</h5>
3245<pre>
Reid Spencerd4622352005-04-26 20:41:16 +00003246 declare void %llvm.memmove(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3247 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00003248</pre>
3249
3250<h5>Overview:</h5>
3251
3252<p>
3253The '<tt>llvm.memmove</tt>' intrinsic moves a block of memory from the source
3254location to the destination location. It is similar to the '<tt>llvm.memcpy</tt>'
3255intrinsic but allows the two memory locations to overlap.
3256</p>
3257
3258<p>
3259Note that, unlike the standard libc function, the <tt>llvm.memmove</tt> intrinsic
3260does not return a value, and takes an extra alignment argument.
3261</p>
3262
3263<h5>Arguments:</h5>
3264
3265<p>
3266The first argument is a pointer to the destination, the second is a pointer to
3267the source. The third argument is an (arbitrarily sized) integer argument
3268specifying the number of bytes to copy, and the fourth argument is the alignment
3269of the source and destination locations.
3270</p>
3271
Chris Lattner3301ced2004-02-12 21:18:15 +00003272<p>
3273If the call to this intrinisic has an alignment value that is not 0 or 1, then
3274the caller guarantees that the size of the copy is a multiple of the alignment
3275and that both the source and destination pointers are aligned to that boundary.
3276</p>
3277
Chris Lattner0eb51b42004-02-12 18:10:10 +00003278<h5>Semantics:</h5>
3279
3280<p>
3281The '<tt>llvm.memmove</tt>' intrinsic copies a block of memory from the source
3282location to the destination location, which may overlap. It
3283copies "len" bytes of memory over. If the argument is known to be aligned to
3284some boundary, this can be specified as the fourth argument, otherwise it should
3285be set to 0 or 1.
3286</p>
3287</div>
3288
Chris Lattner8ff75902004-01-06 05:31:32 +00003289
Chris Lattner10610642004-02-14 04:08:35 +00003290<!-- _______________________________________________________________________ -->
3291<div class="doc_subsubsection">
3292 <a name="i_memset">'<tt>llvm.memset</tt>' Intrinsic</a>
3293</div>
3294
3295<div class="doc_text">
3296
3297<h5>Syntax:</h5>
3298<pre>
Reid Spencerd4622352005-04-26 20:41:16 +00003299 declare void %llvm.memset(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
3300 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003301</pre>
3302
3303<h5>Overview:</h5>
3304
3305<p>
3306The '<tt>llvm.memset</tt>' intrinsic fills a block of memory with a particular
3307byte value.
3308</p>
3309
3310<p>
3311Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
3312does not return a value, and takes an extra alignment argument.
3313</p>
3314
3315<h5>Arguments:</h5>
3316
3317<p>
3318The first argument is a pointer to the destination to fill, the second is the
3319byte value to fill it with, the third argument is an (arbitrarily sized) integer
3320argument specifying the number of bytes to fill, and the fourth argument is the
3321known alignment of destination location.
3322</p>
3323
3324<p>
3325If the call to this intrinisic has an alignment value that is not 0 or 1, then
3326the caller guarantees that the size of the copy is a multiple of the alignment
3327and that the destination pointer is aligned to that boundary.
3328</p>
3329
3330<h5>Semantics:</h5>
3331
3332<p>
3333The '<tt>llvm.memset</tt>' intrinsic fills "len" bytes of memory starting at the
3334destination location. If the argument is known to be aligned to some boundary,
3335this can be specified as the fourth argument, otherwise it should be set to 0 or
33361.
3337</p>
3338</div>
3339
3340
Chris Lattner32006282004-06-11 02:28:03 +00003341<!-- _______________________________________________________________________ -->
3342<div class="doc_subsubsection">
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003343 <a name="i_isunordered">'<tt>llvm.isunordered</tt>' Intrinsic</a>
3344</div>
3345
3346<div class="doc_text">
3347
3348<h5>Syntax:</h5>
3349<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003350 declare bool %llvm.isunordered(&lt;float or double&gt; Val1, &lt;float or double&gt; Val2)
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003351</pre>
3352
3353<h5>Overview:</h5>
3354
3355<p>
3356The '<tt>llvm.isunordered</tt>' intrinsic returns true if either or both of the
3357specified floating point values is a NAN.
3358</p>
3359
3360<h5>Arguments:</h5>
3361
3362<p>
3363The arguments are floating point numbers of the same type.
3364</p>
3365
3366<h5>Semantics:</h5>
3367
3368<p>
3369If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
3370false.
3371</p>
3372</div>
3373
3374
Chris Lattnera4d74142005-07-21 01:29:16 +00003375<!-- _______________________________________________________________________ -->
3376<div class="doc_subsubsection">
3377 <a name="i_sqrt">'<tt>llvm.sqrt</tt>' Intrinsic</a>
3378</div>
3379
3380<div class="doc_text">
3381
3382<h5>Syntax:</h5>
3383<pre>
3384 declare &lt;float or double&gt; %llvm.sqrt(&lt;float or double&gt; Val)
3385</pre>
3386
3387<h5>Overview:</h5>
3388
3389<p>
3390The '<tt>llvm.sqrt</tt>' intrinsic returns the sqrt of the specified operand,
3391returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
3392<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
3393negative numbers (which allows for better optimization).
3394</p>
3395
3396<h5>Arguments:</h5>
3397
3398<p>
3399The argument and return value are floating point numbers of the same type.
3400</p>
3401
3402<h5>Semantics:</h5>
3403
3404<p>
3405This function returns the sqrt of the specified operand if it is a positive
3406floating point number.
3407</p>
3408</div>
3409
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003410<!-- ======================================================================= -->
3411<div class="doc_subsection">
3412 <a name="int_count">Bit Counting Intrinsics</a>
3413</div>
3414
3415<div class="doc_text">
3416<p>
3417LLVM provides intrinsics for a few important bit counting operations.
3418These allow efficient code generation for some algorithms.
3419</p>
3420
3421</div>
3422
3423<!-- _______________________________________________________________________ -->
3424<div class="doc_subsubsection">
3425 <a name="int_ctpop">'<tt>llvm.ctpop</tt>' Intrinsic</a>
3426</div>
3427
3428<div class="doc_text">
3429
3430<h5>Syntax:</h5>
3431<pre>
3432 declare int %llvm.ctpop(int &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003433</pre>
3434
3435<h5>Overview:</h5>
3436
3437<p>
3438The '<tt>llvm.ctpop</tt>' intrinsic counts the number of ones in a variable.
3439</p>
3440
3441<h5>Arguments:</h5>
3442
3443<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003444The only argument is the value to be counted. The argument may be of any
3445integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003446</p>
3447
3448<h5>Semantics:</h5>
3449
3450<p>
3451The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
3452</p>
3453</div>
3454
3455<!-- _______________________________________________________________________ -->
3456<div class="doc_subsubsection">
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003457 <a name="int_ctlz">'<tt>llvm.ctlz</tt>' Intrinsic</a>
3458</div>
3459
3460<div class="doc_text">
3461
3462<h5>Syntax:</h5>
3463<pre>
3464 declare int %llvm.ctlz(int &lt;src&gt;)
3465
3466</pre>
3467
3468<h5>Overview:</h5>
3469
3470<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003471The '<tt>llvm.ctlz</tt>' intrinsic counts the number of leading zeros in a
3472variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003473</p>
3474
3475<h5>Arguments:</h5>
3476
3477<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003478The only argument is the value to be counted. The argument may be of any
3479integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003480</p>
3481
3482<h5>Semantics:</h5>
3483
3484<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00003485The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
3486in a variable. If the src == 0 then the result is the size in bits of the type
3487of src. For example, <tt>llvm.cttz(int 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003488</p>
3489</div>
Chris Lattner32006282004-06-11 02:28:03 +00003490
3491
Chris Lattnereff29ab2005-05-15 19:39:26 +00003492
3493<!-- _______________________________________________________________________ -->
3494<div class="doc_subsubsection">
3495 <a name="int_cttz">'<tt>llvm.cttz</tt>' Intrinsic</a>
3496</div>
3497
3498<div class="doc_text">
3499
3500<h5>Syntax:</h5>
3501<pre>
3502 declare int %llvm.cttz(int &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +00003503</pre>
3504
3505<h5>Overview:</h5>
3506
3507<p>
3508The '<tt>llvm.cttz</tt>' intrinsic counts the number of trailing zeros.
3509</p>
3510
3511<h5>Arguments:</h5>
3512
3513<p>
3514The only argument is the value to be counted. The argument may be of any
3515integer type. The return type must match the argument type.
3516</p>
3517
3518<h5>Semantics:</h5>
3519
3520<p>
3521The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
3522in a variable. If the src == 0 then the result is the size in bits of the type
3523of src. For example, <tt>llvm.cttz(2) = 1</tt>.
3524</p>
3525</div>
3526
Chris Lattner8ff75902004-01-06 05:31:32 +00003527<!-- ======================================================================= -->
3528<div class="doc_subsection">
3529 <a name="int_debugger">Debugger Intrinsics</a>
3530</div>
3531
3532<div class="doc_text">
3533<p>
3534The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
3535are described in the <a
3536href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
3537Debugging</a> document.
3538</p>
3539</div>
3540
3541
Chris Lattner00950542001-06-06 20:29:01 +00003542<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00003543<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00003544<address>
3545 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3546 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3547 <a href="http://validator.w3.org/check/referer"><img
3548 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3549
3550 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
3551 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
3552 Last modified: $Date$
3553</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003554</body>
3555</html>