blob: ab0d66d96ed844329503a5b1ddcc74354021ea55 [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>
Chris Lattner4e9aba72006-01-23 23:23:47 +000026 <li><a href="#functionstructure">Functions</a></li>
27 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000028 </ol>
29 </li>
Chris Lattner00950542001-06-06 20:29:01 +000030 <li><a href="#typesystem">Type System</a>
31 <ol>
Robert Bocchino7b81c752006-02-17 21:18:08 +000032 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +000033 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +000034 <li><a href="#t_classifications">Type Classifications</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000035 </ol>
36 </li>
Chris Lattner00950542001-06-06 20:29:01 +000037 <li><a href="#t_derived">Derived Types</a>
38 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000039 <li><a href="#t_array">Array Type</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000040 <li><a href="#t_function">Function Type</a></li>
41 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000042 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattnera58561b2004-08-12 19:12:28 +000043 <li><a href="#t_packed">Packed Type</a></li>
Chris Lattner69c11bb2005-04-25 17:34:15 +000044 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000045 </ol>
46 </li>
47 </ol>
48 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +000049 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +000050 <ol>
51 <li><a href="#simpleconstants">Simple Constants</a>
52 <li><a href="#aggregateconstants">Aggregate Constants</a>
53 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
54 <li><a href="#undefvalues">Undefined Values</a>
55 <li><a href="#constantexprs">Constant Expressions</a>
56 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +000057 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +000058 <li><a href="#othervalues">Other Values</a>
59 <ol>
60 <li><a href="#inlineasm">Inline Assembler Expressions</a>
61 </ol>
62 </li>
Chris Lattner00950542001-06-06 20:29:01 +000063 <li><a href="#instref">Instruction Reference</a>
64 <ol>
65 <li><a href="#terminators">Terminator Instructions</a>
66 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000067 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
68 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000069 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
70 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000071 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +000072 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000073 </ol>
74 </li>
Chris Lattner00950542001-06-06 20:29:01 +000075 <li><a href="#binaryops">Binary Operations</a>
76 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000077 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
78 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
79 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
80 <li><a href="#i_div">'<tt>div</tt>' Instruction</a></li>
81 <li><a href="#i_rem">'<tt>rem</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000082 <li><a href="#i_setcc">'<tt>set<i>cc</i></tt>' Instructions</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000083 </ol>
84 </li>
Chris Lattner00950542001-06-06 20:29:01 +000085 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
86 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +000087 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000088 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000089 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
90 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
91 <li><a href="#i_shr">'<tt>shr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000092 </ol>
93 </li>
Chris Lattner00950542001-06-06 20:29:01 +000094 <li><a href="#memoryops">Memory Access Operations</a>
95 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000096 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
97 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
98 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +000099 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
100 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
101 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000102 </ol>
103 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000104 <li><a href="#otherops">Other Operations</a>
105 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000106 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000107 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000108 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Robert Bocchino8949de22006-03-14 20:55:28 +0000109 <li><a href="#i_vsetint">'<tt>vsetint</tt>' Instruction</a></li>
110 <li><a href="#i_vsetfp">'<tt>vsetfp</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000111 <li><a href="#i_vselect">'<tt>vselect</tt>' Instruction</a></li>
Robert Bocchino3a558662006-01-05 17:37:02 +0000112 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
Robert Bocchino05ccd702006-01-15 20:48:27 +0000113 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000114 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000115 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000116 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000117 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000118 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000119 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000120 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000121 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000122 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
123 <ol>
124 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
125 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
126 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
127 </ol>
128 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000129 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
130 <ol>
131 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
132 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
133 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
134 </ol>
135 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000136 <li><a href="#int_codegen">Code Generator Intrinsics</a>
137 <ol>
138 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
139 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
Chris Lattner57e1f392006-01-13 02:03:13 +0000140 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
141 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000142 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000143 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000144 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000145 </ol>
146 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000147 <li><a href="#int_libc">Standard C Library Intrinsics</a>
148 <ol>
Chris Lattner5b310c32006-03-03 00:07:20 +0000149 <li><a href="#i_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
150 <li><a href="#i_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
151 <li><a href="#i_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000152 <li><a href="#i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a></li>
153 <li><a href="#i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
Chris Lattnera4d74142005-07-21 01:29:16 +0000154
Chris Lattner33aec9e2004-02-12 17:01:32 +0000155 </ol>
156 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000157 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000158 <ol>
Nate Begeman7e36c472006-01-13 23:26:38 +0000159 <li><a href="#i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000160 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
161 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
162 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000163 </ol>
164 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000165 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000166 </ol>
167 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000168</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000169
170<div class="doc_author">
171 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
172 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000173</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000174
Chris Lattner00950542001-06-06 20:29:01 +0000175<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000176<div class="doc_section"> <a name="abstract">Abstract </a></div>
177<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000178
Misha Brukman9d0919f2003-11-08 01:05:38 +0000179<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000180<p>This document is a reference manual for the LLVM assembly language.
181LLVM is an SSA based representation that provides type safety,
182low-level operations, flexibility, and the capability of representing
183'all' high-level languages cleanly. It is the common code
184representation used throughout all phases of the LLVM compilation
185strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000186</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000187
Chris Lattner00950542001-06-06 20:29:01 +0000188<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000189<div class="doc_section"> <a name="introduction">Introduction</a> </div>
190<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000191
Misha Brukman9d0919f2003-11-08 01:05:38 +0000192<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000193
Chris Lattner261efe92003-11-25 01:02:51 +0000194<p>The LLVM code representation is designed to be used in three
195different forms: as an in-memory compiler IR, as an on-disk bytecode
196representation (suitable for fast loading by a Just-In-Time compiler),
197and as a human readable assembly language representation. This allows
198LLVM to provide a powerful intermediate representation for efficient
199compiler transformations and analysis, while providing a natural means
200to debug and visualize the transformations. The three different forms
201of LLVM are all equivalent. This document describes the human readable
202representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000203
John Criswellc1f786c2005-05-13 22:25:59 +0000204<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000205while being expressive, typed, and extensible at the same time. It
206aims to be a "universal IR" of sorts, by being at a low enough level
207that high-level ideas may be cleanly mapped to it (similar to how
208microprocessors are "universal IR's", allowing many source languages to
209be mapped to them). By providing type information, LLVM can be used as
210the target of optimizations: for example, through pointer analysis, it
211can be proven that a C automatic variable is never accessed outside of
212the current function... allowing it to be promoted to a simple SSA
213value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000214
Misha Brukman9d0919f2003-11-08 01:05:38 +0000215</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000216
Chris Lattner00950542001-06-06 20:29:01 +0000217<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000218<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000219
Misha Brukman9d0919f2003-11-08 01:05:38 +0000220<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000221
Chris Lattner261efe92003-11-25 01:02:51 +0000222<p>It is important to note that this document describes 'well formed'
223LLVM assembly language. There is a difference between what the parser
224accepts and what is considered 'well formed'. For example, the
225following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000226
227<pre>
228 %x = <a href="#i_add">add</a> int 1, %x
229</pre>
230
Chris Lattner261efe92003-11-25 01:02:51 +0000231<p>...because the definition of <tt>%x</tt> does not dominate all of
232its uses. The LLVM infrastructure provides a verification pass that may
233be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000234automatically run by the parser after parsing input assembly and by
Chris Lattner261efe92003-11-25 01:02:51 +0000235the optimizer before it outputs bytecode. The violations pointed out
236by the verifier pass indicate bugs in transformation passes or input to
237the parser.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000238
Chris Lattner261efe92003-11-25 01:02:51 +0000239<!-- Describe the typesetting conventions here. --> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000240
Chris Lattner00950542001-06-06 20:29:01 +0000241<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000242<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000243<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000244
Misha Brukman9d0919f2003-11-08 01:05:38 +0000245<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000246
Chris Lattner261efe92003-11-25 01:02:51 +0000247<p>LLVM uses three different forms of identifiers, for different
248purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000249
Chris Lattner00950542001-06-06 20:29:01 +0000250<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000251 <li>Named values are represented as a string of characters with a '%' prefix.
252 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
253 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
254 Identifiers which require other characters in their names can be surrounded
255 with quotes. In this way, anything except a <tt>"</tt> character can be used
256 in a name.</li>
257
258 <li>Unnamed values are represented as an unsigned numeric value with a '%'
259 prefix. For example, %12, %2, %44.</li>
260
Reid Spencercc16dc32004-12-09 18:02:53 +0000261 <li>Constants, which are described in a <a href="#constants">section about
262 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000263</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000264
265<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
266don't need to worry about name clashes with reserved words, and the set of
267reserved words may be expanded in the future without penalty. Additionally,
268unnamed identifiers allow a compiler to quickly come up with a temporary
269variable without having to avoid symbol table conflicts.</p>
270
Chris Lattner261efe92003-11-25 01:02:51 +0000271<p>Reserved words in LLVM are very similar to reserved words in other
272languages. There are keywords for different opcodes ('<tt><a
Chris Lattnere5d947b2004-12-09 16:36:40 +0000273href="#i_add">add</a></tt>', '<tt><a href="#i_cast">cast</a></tt>', '<tt><a
274href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
275href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
276and others. These reserved words cannot conflict with variable names, because
277none of them start with a '%' character.</p>
278
279<p>Here is an example of LLVM code to multiply the integer variable
280'<tt>%X</tt>' by 8:</p>
281
Misha Brukman9d0919f2003-11-08 01:05:38 +0000282<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000283
284<pre>
285 %result = <a href="#i_mul">mul</a> uint %X, 8
286</pre>
287
Misha Brukman9d0919f2003-11-08 01:05:38 +0000288<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000289
290<pre>
291 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
292</pre>
293
Misha Brukman9d0919f2003-11-08 01:05:38 +0000294<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000295
296<pre>
297 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
298 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
299 %result = <a href="#i_add">add</a> uint %1, %1
300</pre>
301
Chris Lattner261efe92003-11-25 01:02:51 +0000302<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
303important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000304
Chris Lattner00950542001-06-06 20:29:01 +0000305<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000306
307 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
308 line.</li>
309
310 <li>Unnamed temporaries are created when the result of a computation is not
311 assigned to a named value.</li>
312
Misha Brukman9d0919f2003-11-08 01:05:38 +0000313 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000314
Misha Brukman9d0919f2003-11-08 01:05:38 +0000315</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000316
John Criswelle4c57cc2005-05-12 16:52:32 +0000317<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000318demonstrating instructions, we will follow an instruction with a comment that
319defines the type and name of value produced. Comments are shown in italic
320text.</p>
321
Misha Brukman9d0919f2003-11-08 01:05:38 +0000322</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000323
324<!-- *********************************************************************** -->
325<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
326<!-- *********************************************************************** -->
327
328<!-- ======================================================================= -->
329<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
330</div>
331
332<div class="doc_text">
333
334<p>LLVM programs are composed of "Module"s, each of which is a
335translation unit of the input programs. Each module consists of
336functions, global variables, and symbol table entries. Modules may be
337combined together with the LLVM linker, which merges function (and
338global variable) definitions, resolves forward declarations, and merges
339symbol table entries. Here is an example of the "hello world" module:</p>
340
341<pre><i>; Declare the string constant as a global constant...</i>
342<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
343 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
344
345<i>; External declaration of the puts function</i>
346<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
347
348<i>; Definition of main function</i>
349int %main() { <i>; int()* </i>
350 <i>; Convert [13x sbyte]* to sbyte *...</i>
351 %cast210 = <a
352 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
353
354 <i>; Call puts function to write out the string to stdout...</i>
355 <a
356 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
357 <a
358 href="#i_ret">ret</a> int 0<br>}<br></pre>
359
360<p>This example is made up of a <a href="#globalvars">global variable</a>
361named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
362function, and a <a href="#functionstructure">function definition</a>
363for "<tt>main</tt>".</p>
364
Chris Lattnere5d947b2004-12-09 16:36:40 +0000365<p>In general, a module is made up of a list of global values,
366where both functions and global variables are global values. Global values are
367represented by a pointer to a memory location (in this case, a pointer to an
368array of char, and a pointer to a function), and have one of the following <a
369href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000370
Chris Lattnere5d947b2004-12-09 16:36:40 +0000371</div>
372
373<!-- ======================================================================= -->
374<div class="doc_subsection">
375 <a name="linkage">Linkage Types</a>
376</div>
377
378<div class="doc_text">
379
380<p>
381All Global Variables and Functions have one of the following types of linkage:
382</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000383
384<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000385
Chris Lattnerfa730212004-12-09 16:11:40 +0000386 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000387
388 <dd>Global values with internal linkage are only directly accessible by
389 objects in the current module. In particular, linking code into a module with
390 an internal global value may cause the internal to be renamed as necessary to
391 avoid collisions. Because the symbol is internal to the module, all
392 references can be updated. This corresponds to the notion of the
393 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattnerfa730212004-12-09 16:11:40 +0000394 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000395
Chris Lattnerfa730212004-12-09 16:11:40 +0000396 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000397
398 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
399 the twist that linking together two modules defining the same
400 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
401 is typically used to implement inline functions. Unreferenced
402 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000403 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000404
Chris Lattnerfa730212004-12-09 16:11:40 +0000405 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000406
407 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
408 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
409 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000410 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000411
Chris Lattnerfa730212004-12-09 16:11:40 +0000412 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000413
414 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
415 pointer to array type. When two global variables with appending linkage are
416 linked together, the two global arrays are appended together. This is the
417 LLVM, typesafe, equivalent of having the system linker append together
418 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000419 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000420
Chris Lattnerfa730212004-12-09 16:11:40 +0000421 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000422
423 <dd>If none of the above identifiers are used, the global is externally
424 visible, meaning that it participates in linkage and can be used to resolve
425 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000426 </dd>
427</dl>
428
Chris Lattnerfa730212004-12-09 16:11:40 +0000429<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
430variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
431variable and was linked with this one, one of the two would be renamed,
432preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
433external (i.e., lacking any linkage declarations), they are accessible
434outside of the current module. It is illegal for a function <i>declaration</i>
435to have any linkage type other than "externally visible".</a></p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000436
Chris Lattnerfa730212004-12-09 16:11:40 +0000437</div>
438
439<!-- ======================================================================= -->
440<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000441 <a name="callingconv">Calling Conventions</a>
442</div>
443
444<div class="doc_text">
445
446<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
447and <a href="#i_invoke">invokes</a> can all have an optional calling convention
448specified for the call. The calling convention of any pair of dynamic
449caller/callee must match, or the behavior of the program is undefined. The
450following calling conventions are supported by LLVM, and more may be added in
451the future:</p>
452
453<dl>
454 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
455
456 <dd>This calling convention (the default if no other calling convention is
457 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000458 supports varargs function calls and tolerates some mismatch in the declared
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000459 prototype and implemented declaration of the function (as does normal C).
460 </dd>
461
462 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
463
464 <dd>This calling convention attempts to make calls as fast as possible
465 (e.g. by passing things in registers). This calling convention allows the
466 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000467 without having to conform to an externally specified ABI. Implementations of
468 this convention should allow arbitrary tail call optimization to be supported.
469 This calling convention does not support varargs and requires the prototype of
470 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000471 </dd>
472
473 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
474
475 <dd>This calling convention attempts to make code in the caller as efficient
476 as possible under the assumption that the call is not commonly executed. As
477 such, these calls often preserve all registers so that the call does not break
478 any live ranges in the caller side. This calling convention does not support
479 varargs and requires the prototype of all callees to exactly match the
480 prototype of the function definition.
481 </dd>
482
Chris Lattnercfe6b372005-05-07 01:46:40 +0000483 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000484
485 <dd>Any calling convention may be specified by number, allowing
486 target-specific calling conventions to be used. Target specific calling
487 conventions start at 64.
488 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000489</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000490
491<p>More calling conventions can be added/defined on an as-needed basis, to
492support pascal conventions or any other well-known target-independent
493convention.</p>
494
495</div>
496
497<!-- ======================================================================= -->
498<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000499 <a name="globalvars">Global Variables</a>
500</div>
501
502<div class="doc_text">
503
Chris Lattner3689a342005-02-12 19:30:21 +0000504<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000505instead of run-time. Global variables may optionally be initialized, may have
506an explicit section to be placed in, and may
Chris Lattner2cbdc452005-11-06 08:02:57 +0000507have an optional explicit alignment specified. A
John Criswell0ec250c2005-10-24 16:17:18 +0000508variable may be defined as a global "constant," which indicates that the
Chris Lattner3689a342005-02-12 19:30:21 +0000509contents of the variable will <b>never</b> be modified (enabling better
510optimization, allowing the global data to be placed in the read-only section of
511an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000512cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000513
514<p>
515LLVM explicitly allows <em>declarations</em> of global variables to be marked
516constant, even if the final definition of the global is not. This capability
517can be used to enable slightly better optimization of the program, but requires
518the language definition to guarantee that optimizations based on the
519'constantness' are valid for the translation units that do not include the
520definition.
521</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000522
523<p>As SSA values, global variables define pointer values that are in
524scope (i.e. they dominate) all basic blocks in the program. Global
525variables always define a pointer to their "content" type because they
526describe a region of memory, and all memory objects in LLVM are
527accessed through pointers.</p>
528
Chris Lattner88f6c462005-11-12 00:45:07 +0000529<p>LLVM allows an explicit section to be specified for globals. If the target
530supports it, it will emit globals to the section specified.</p>
531
Chris Lattner2cbdc452005-11-06 08:02:57 +0000532<p>An explicit alignment may be specified for a global. If not present, or if
533the alignment is set to zero, the alignment of the global is set by the target
534to whatever it feels convenient. If an explicit alignment is specified, the
535global is forced to have at least that much alignment. All alignments must be
536a power of 2.</p>
537
Chris Lattnerfa730212004-12-09 16:11:40 +0000538</div>
539
540
541<!-- ======================================================================= -->
542<div class="doc_subsection">
543 <a name="functionstructure">Functions</a>
544</div>
545
546<div class="doc_text">
547
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000548<p>LLVM function definitions consist of an optional <a href="#linkage">linkage
549type</a>, an optional <a href="#callingconv">calling convention</a>, a return
Chris Lattner88f6c462005-11-12 00:45:07 +0000550type, a function name, a (possibly empty) argument list, an optional section,
551an optional alignment, an opening curly brace,
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000552a list of basic blocks, and a closing curly brace. LLVM function declarations
553are defined with the "<tt>declare</tt>" keyword, an optional <a
Chris Lattner2cbdc452005-11-06 08:02:57 +0000554href="#callingconv">calling convention</a>, a return type, a function name,
555a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000556
557<p>A function definition contains a list of basic blocks, forming the CFG for
558the function. Each basic block may optionally start with a label (giving the
559basic block a symbol table entry), contains a list of instructions, and ends
560with a <a href="#terminators">terminator</a> instruction (such as a branch or
561function return).</p>
562
John Criswelle4c57cc2005-05-12 16:52:32 +0000563<p>The first basic block in a program is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000564executed on entrance to the function, and it is not allowed to have predecessor
565basic blocks (i.e. there can not be any branches to the entry block of a
566function). Because the block can have no predecessors, it also cannot have any
567<a href="#i_phi">PHI nodes</a>.</p>
568
569<p>LLVM functions are identified by their name and type signature. Hence, two
570functions with the same name but different parameter lists or return values are
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000571considered different functions, and LLVM will resolve references to each
Chris Lattnerfa730212004-12-09 16:11:40 +0000572appropriately.</p>
573
Chris Lattner88f6c462005-11-12 00:45:07 +0000574<p>LLVM allows an explicit section to be specified for functions. If the target
575supports it, it will emit functions to the section specified.</p>
576
Chris Lattner2cbdc452005-11-06 08:02:57 +0000577<p>An explicit alignment may be specified for a function. If not present, or if
578the alignment is set to zero, the alignment of the function is set by the target
579to whatever it feels convenient. If an explicit alignment is specified, the
580function is forced to have at least that much alignment. All alignments must be
581a power of 2.</p>
582
Chris Lattnerfa730212004-12-09 16:11:40 +0000583</div>
584
Chris Lattner4e9aba72006-01-23 23:23:47 +0000585<!-- ======================================================================= -->
586<div class="doc_subsection">
587 <a name="moduleasm">Module-Level Inline Assembly</a></li>
588</div>
589
590<div class="doc_text">
591<p>
592Modules may contain "module-level inline asm" blocks, which corresponds to the
593GCC "file scope inline asm" blocks. These blocks are internally concatenated by
594LLVM and treated as a single unit, but may be separated in the .ll file if
595desired. The syntax is very simple:
596</p>
597
598<div class="doc_code"><pre>
Chris Lattner52599e12006-01-24 00:37:20 +0000599 module asm "inline asm code goes here"
600 module asm "more can go here"
Chris Lattner4e9aba72006-01-23 23:23:47 +0000601</pre></div>
602
603<p>The strings can contain any character by escaping non-printable characters.
604 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
605 for the number.
606</p>
607
608<p>
609 The inline asm code is simply printed to the machine code .s file when
610 assembly code is generated.
611</p>
612</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000613
614
Chris Lattner00950542001-06-06 20:29:01 +0000615<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000616<div class="doc_section"> <a name="typesystem">Type System</a> </div>
617<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000618
Misha Brukman9d0919f2003-11-08 01:05:38 +0000619<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000620
Misha Brukman9d0919f2003-11-08 01:05:38 +0000621<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000622intermediate representation. Being typed enables a number of
623optimizations to be performed on the IR directly, without having to do
624extra analyses on the side before the transformation. A strong type
625system makes it easier to read the generated code and enables novel
626analyses and transformations that are not feasible to perform on normal
627three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000628
629</div>
630
Chris Lattner00950542001-06-06 20:29:01 +0000631<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000632<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000633<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000634<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000635system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000636
Reid Spencerd3f876c2004-11-01 08:19:36 +0000637<table class="layout">
638 <tr class="layout">
639 <td class="left">
640 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000641 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000642 <tr><th>Type</th><th>Description</th></tr>
643 <tr><td><tt>void</tt></td><td>No value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000644 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
645 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
646 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
647 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
648 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000649 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000650 </tbody>
651 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000652 </td>
653 <td class="right">
654 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000655 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000656 <tr><th>Type</th><th>Description</th></tr>
657 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000658 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
659 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
660 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
661 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
662 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000663 </tbody>
664 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000665 </td>
666 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000667</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000668</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000669
Chris Lattner00950542001-06-06 20:29:01 +0000670<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000671<div class="doc_subsubsection"> <a name="t_classifications">Type
672Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000673<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000674<p>These different primitive types fall into a few useful
675classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000676
677<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000678 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000679 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000680 <tr>
681 <td><a name="t_signed">signed</a></td>
682 <td><tt>sbyte, short, int, long, float, double</tt></td>
683 </tr>
684 <tr>
685 <td><a name="t_unsigned">unsigned</a></td>
686 <td><tt>ubyte, ushort, uint, ulong</tt></td>
687 </tr>
688 <tr>
689 <td><a name="t_integer">integer</a></td>
690 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
691 </tr>
692 <tr>
693 <td><a name="t_integral">integral</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000694 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
695 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000696 </tr>
697 <tr>
698 <td><a name="t_floating">floating point</a></td>
699 <td><tt>float, double</tt></td>
700 </tr>
701 <tr>
702 <td><a name="t_firstclass">first class</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000703 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
704 float, double, <a href="#t_pointer">pointer</a>,
705 <a href="#t_packed">packed</a></tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000706 </tr>
707 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000708</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000709
Chris Lattner261efe92003-11-25 01:02:51 +0000710<p>The <a href="#t_firstclass">first class</a> types are perhaps the
711most important. Values of these types are the only ones which can be
712produced by instructions, passed as arguments, or used as operands to
713instructions. This means that all structures and arrays must be
714manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000715</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000716
Chris Lattner00950542001-06-06 20:29:01 +0000717<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000718<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000719
Misha Brukman9d0919f2003-11-08 01:05:38 +0000720<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000721
Chris Lattner261efe92003-11-25 01:02:51 +0000722<p>The real power in LLVM comes from the derived types in the system.
723This is what allows a programmer to represent arrays, functions,
724pointers, and other useful types. Note that these derived types may be
725recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000726
Misha Brukman9d0919f2003-11-08 01:05:38 +0000727</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000728
Chris Lattner00950542001-06-06 20:29:01 +0000729<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000730<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000731
Misha Brukman9d0919f2003-11-08 01:05:38 +0000732<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000733
Chris Lattner00950542001-06-06 20:29:01 +0000734<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000735
Misha Brukman9d0919f2003-11-08 01:05:38 +0000736<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000737sequentially in memory. The array type requires a size (number of
738elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000739
Chris Lattner7faa8832002-04-14 06:13:44 +0000740<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000741
742<pre>
743 [&lt;# elements&gt; x &lt;elementtype&gt;]
744</pre>
745
John Criswelle4c57cc2005-05-12 16:52:32 +0000746<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000747be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000748
Chris Lattner7faa8832002-04-14 06:13:44 +0000749<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000750<table class="layout">
751 <tr class="layout">
752 <td class="left">
753 <tt>[40 x int ]</tt><br/>
754 <tt>[41 x int ]</tt><br/>
755 <tt>[40 x uint]</tt><br/>
756 </td>
757 <td class="left">
758 Array of 40 integer values.<br/>
759 Array of 41 integer values.<br/>
760 Array of 40 unsigned integer values.<br/>
761 </td>
762 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000763</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000764<p>Here are some examples of multidimensional arrays:</p>
765<table class="layout">
766 <tr class="layout">
767 <td class="left">
768 <tt>[3 x [4 x int]]</tt><br/>
769 <tt>[12 x [10 x float]]</tt><br/>
770 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
771 </td>
772 <td class="left">
John Criswellc1f786c2005-05-13 22:25:59 +0000773 3x4 array of integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000774 12x10 array of single precision floating point values.<br/>
775 2x3x4 array of unsigned integer values.<br/>
776 </td>
777 </tr>
778</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000779
John Criswell0ec250c2005-10-24 16:17:18 +0000780<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
781length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000782LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
783As a special case, however, zero length arrays are recognized to be variable
784length. This allows implementation of 'pascal style arrays' with the LLVM
785type "{ int, [0 x float]}", for example.</p>
786
Misha Brukman9d0919f2003-11-08 01:05:38 +0000787</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000788
Chris Lattner00950542001-06-06 20:29:01 +0000789<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000790<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000791<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000792<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000793<p>The function type can be thought of as a function signature. It
794consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000795Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000796(which are structures of pointers to functions), for indirect function
797calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000798<p>
799The return type of a function type cannot be an aggregate type.
800</p>
Chris Lattner00950542001-06-06 20:29:01 +0000801<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000802<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000803<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +0000804specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000805which indicates that the function takes a variable number of arguments.
806Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000807 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000808<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000809<table class="layout">
810 <tr class="layout">
811 <td class="left">
812 <tt>int (int)</tt> <br/>
813 <tt>float (int, int *) *</tt><br/>
814 <tt>int (sbyte *, ...)</tt><br/>
815 </td>
816 <td class="left">
817 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
818 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukmanc24b7582004-08-12 20:16:08 +0000819 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
Reid Spencerd3f876c2004-11-01 08:19:36 +0000820 returning <tt>float</tt>.<br/>
821 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
822 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
823 the signature for <tt>printf</tt> in LLVM.<br/>
824 </td>
825 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000826</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000827
Misha Brukman9d0919f2003-11-08 01:05:38 +0000828</div>
Chris Lattner00950542001-06-06 20:29:01 +0000829<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000830<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000831<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000832<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000833<p>The structure type is used to represent a collection of data members
834together in memory. The packing of the field types is defined to match
835the ABI of the underlying processor. The elements of a structure may
836be any type that has a size.</p>
837<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
838and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
839field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
840instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000841<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000842<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000843<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000844<table class="layout">
845 <tr class="layout">
846 <td class="left">
847 <tt>{ int, int, int }</tt><br/>
848 <tt>{ float, int (int) * }</tt><br/>
849 </td>
850 <td class="left">
851 a triple of three <tt>int</tt> values<br/>
852 A pair, where the first element is a <tt>float</tt> and the second element
853 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
854 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
855 </td>
856 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000857</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000858</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000859
Chris Lattner00950542001-06-06 20:29:01 +0000860<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000861<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000862<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000863<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000864<p>As in many languages, the pointer type represents a pointer or
865reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000866<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000867<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000868<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000869<table class="layout">
870 <tr class="layout">
871 <td class="left">
872 <tt>[4x int]*</tt><br/>
873 <tt>int (int *) *</tt><br/>
874 </td>
875 <td class="left">
876 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
877 four <tt>int</tt> values<br/>
878 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattnera977c482005-02-19 02:22:14 +0000879 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerd3f876c2004-11-01 08:19:36 +0000880 <tt>int</tt>.<br/>
881 </td>
882 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000883</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000884</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000885
Chris Lattnera58561b2004-08-12 19:12:28 +0000886<!-- _______________________________________________________________________ -->
887<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000888<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000889
Chris Lattnera58561b2004-08-12 19:12:28 +0000890<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000891
Chris Lattnera58561b2004-08-12 19:12:28 +0000892<p>A packed type is a simple derived type that represents a vector
893of elements. Packed types are used when multiple primitive data
894are operated in parallel using a single instruction (SIMD).
895A packed type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000896elements) and an underlying primitive data type. Vectors must have a power
897of two length (1, 2, 4, 8, 16 ...). Packed types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000898considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000899
Chris Lattnera58561b2004-08-12 19:12:28 +0000900<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000901
902<pre>
903 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
904</pre>
905
John Criswellc1f786c2005-05-13 22:25:59 +0000906<p>The number of elements is a constant integer value; elementtype may
Chris Lattnera58561b2004-08-12 19:12:28 +0000907be any integral or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000908
Chris Lattnera58561b2004-08-12 19:12:28 +0000909<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000910
Reid Spencerd3f876c2004-11-01 08:19:36 +0000911<table class="layout">
912 <tr class="layout">
913 <td class="left">
914 <tt>&lt;4 x int&gt;</tt><br/>
915 <tt>&lt;8 x float&gt;</tt><br/>
916 <tt>&lt;2 x uint&gt;</tt><br/>
917 </td>
918 <td class="left">
919 Packed vector of 4 integer values.<br/>
920 Packed vector of 8 floating-point values.<br/>
921 Packed vector of 2 unsigned integer values.<br/>
922 </td>
923 </tr>
924</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000925</div>
926
Chris Lattner69c11bb2005-04-25 17:34:15 +0000927<!-- _______________________________________________________________________ -->
928<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
929<div class="doc_text">
930
931<h5>Overview:</h5>
932
933<p>Opaque types are used to represent unknown types in the system. This
934corresponds (for example) to the C notion of a foward declared structure type.
935In LLVM, opaque types can eventually be resolved to any type (not just a
936structure type).</p>
937
938<h5>Syntax:</h5>
939
940<pre>
941 opaque
942</pre>
943
944<h5>Examples:</h5>
945
946<table class="layout">
947 <tr class="layout">
948 <td class="left">
949 <tt>opaque</tt>
950 </td>
951 <td class="left">
952 An opaque type.<br/>
953 </td>
954 </tr>
955</table>
956</div>
957
958
Chris Lattnerc3f59762004-12-09 17:30:23 +0000959<!-- *********************************************************************** -->
960<div class="doc_section"> <a name="constants">Constants</a> </div>
961<!-- *********************************************************************** -->
962
963<div class="doc_text">
964
965<p>LLVM has several different basic types of constants. This section describes
966them all and their syntax.</p>
967
968</div>
969
970<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000971<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000972
973<div class="doc_text">
974
975<dl>
976 <dt><b>Boolean constants</b></dt>
977
978 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
979 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
980 </dd>
981
982 <dt><b>Integer constants</b></dt>
983
Reid Spencercc16dc32004-12-09 18:02:53 +0000984 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000985 href="#t_integer">integer</a> type. Negative numbers may be used with signed
986 integer types.
987 </dd>
988
989 <dt><b>Floating point constants</b></dt>
990
991 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
992 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000993 notation (see below). Floating point constants must have a <a
994 href="#t_floating">floating point</a> type. </dd>
995
996 <dt><b>Null pointer constants</b></dt>
997
John Criswell9e2485c2004-12-10 15:51:16 +0000998 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +0000999 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1000
1001</dl>
1002
John Criswell9e2485c2004-12-10 15:51:16 +00001003<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +00001004of floating point constants. For example, the form '<tt>double
10050x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
10064.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +00001007(and the only time that they are generated by the disassembler) is when a
1008floating point constant must be emitted but it cannot be represented as a
1009decimal floating point number. For example, NaN's, infinities, and other
1010special values are represented in their IEEE hexadecimal format so that
1011assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001012
1013</div>
1014
1015<!-- ======================================================================= -->
1016<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1017</div>
1018
1019<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001020<p>Aggregate constants arise from aggregation of simple constants
1021and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001022
1023<dl>
1024 <dt><b>Structure constants</b></dt>
1025
1026 <dd>Structure constants are represented with notation similar to structure
1027 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001028 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
1029 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
1030 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +00001031 types of elements must match those specified by the type.
1032 </dd>
1033
1034 <dt><b>Array constants</b></dt>
1035
1036 <dd>Array constants are represented with notation similar to array type
1037 definitions (a comma separated list of elements, surrounded by square brackets
John Criswell9e2485c2004-12-10 15:51:16 +00001038 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +00001039 constants must have <a href="#t_array">array type</a>, and the number and
1040 types of elements must match those specified by the type.
1041 </dd>
1042
1043 <dt><b>Packed constants</b></dt>
1044
1045 <dd>Packed constants are represented with notation similar to packed type
1046 definitions (a comma separated list of elements, surrounded by
John Criswell9e2485c2004-12-10 15:51:16 +00001047 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; int 42,
Chris Lattnerc3f59762004-12-09 17:30:23 +00001048 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
1049 href="#t_packed">packed type</a>, and the number and types of elements must
1050 match those specified by the type.
1051 </dd>
1052
1053 <dt><b>Zero initialization</b></dt>
1054
1055 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1056 value to zero of <em>any</em> type, including scalar and aggregate types.
1057 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +00001058 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +00001059 initializers.
1060 </dd>
1061</dl>
1062
1063</div>
1064
1065<!-- ======================================================================= -->
1066<div class="doc_subsection">
1067 <a name="globalconstants">Global Variable and Function Addresses</a>
1068</div>
1069
1070<div class="doc_text">
1071
1072<p>The addresses of <a href="#globalvars">global variables</a> and <a
1073href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +00001074constants. These constants are explicitly referenced when the <a
1075href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +00001076href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1077file:</p>
1078
1079<pre>
1080 %X = global int 17
1081 %Y = global int 42
1082 %Z = global [2 x int*] [ int* %X, int* %Y ]
1083</pre>
1084
1085</div>
1086
1087<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +00001088<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001089<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +00001090 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +00001091 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +00001092 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001093
Reid Spencer2dc45b82004-12-09 18:13:12 +00001094 <p>Undefined values indicate to the compiler that the program is well defined
1095 no matter what value is used, giving the compiler more freedom to optimize.
1096 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001097</div>
1098
1099<!-- ======================================================================= -->
1100<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1101</div>
1102
1103<div class="doc_text">
1104
1105<p>Constant expressions are used to allow expressions involving other constants
1106to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +00001107href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +00001108that does not have side effects (e.g. load and call are not supported). The
1109following is the syntax for constant expressions:</p>
1110
1111<dl>
1112 <dt><b><tt>cast ( CST to TYPE )</tt></b></dt>
1113
1114 <dd>Cast a constant to another type.</dd>
1115
1116 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1117
1118 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1119 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1120 instruction, the index list may have zero or more indexes, which are required
1121 to make sense for the type of "CSTPTR".</dd>
1122
Robert Bocchino9fbe1452006-01-10 19:31:34 +00001123 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1124
1125 <dd>Perform the <a href="#i_select">select operation</a> on
1126 constants.
1127
1128 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1129
1130 <dd>Perform the <a href="#i_extractelement">extractelement
1131 operation</a> on constants.
1132
Robert Bocchino05ccd702006-01-15 20:48:27 +00001133 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1134
1135 <dd>Perform the <a href="#i_insertelement">insertelement
1136 operation</a> on constants.
1137
Chris Lattnerc3f59762004-12-09 17:30:23 +00001138 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1139
Reid Spencer2dc45b82004-12-09 18:13:12 +00001140 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1141 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +00001142 binary</a> operations. The constraints on operands are the same as those for
1143 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +00001144 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001145</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001146</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00001147
Chris Lattner00950542001-06-06 20:29:01 +00001148<!-- *********************************************************************** -->
Chris Lattnere87d6532006-01-25 23:47:57 +00001149<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1150<!-- *********************************************************************** -->
1151
1152<!-- ======================================================================= -->
1153<div class="doc_subsection">
1154<a name="inlineasm">Inline Assembler Expressions</a>
1155</div>
1156
1157<div class="doc_text">
1158
1159<p>
1160LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1161Module-Level Inline Assembly</a>) through the use of a special value. This
1162value represents the inline assembler as a string (containing the instructions
1163to emit), a list of operand constraints (stored as a string), and a flag that
1164indicates whether or not the inline asm expression has side effects. An example
1165inline assembler expression is:
1166</p>
1167
1168<pre>
1169 int(int) asm "bswap $0", "=r,r"
1170</pre>
1171
1172<p>
1173Inline assembler expressions may <b>only</b> be used as the callee operand of
1174a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1175</p>
1176
1177<pre>
1178 %X = call int asm "<a href="#i_bswap">bswap</a> $0", "=r,r"(int %Y)
1179</pre>
1180
1181<p>
1182Inline asms with side effects not visible in the constraint list must be marked
1183as having side effects. This is done through the use of the
1184'<tt>sideeffect</tt>' keyword, like so:
1185</p>
1186
1187<pre>
1188 call void asm sideeffect "eieio", ""()
1189</pre>
1190
1191<p>TODO: The format of the asm and constraints string still need to be
1192documented here. Constraints on what can be done (e.g. duplication, moving, etc
1193need to be documented).
1194</p>
1195
1196</div>
1197
1198<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00001199<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1200<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00001201
Misha Brukman9d0919f2003-11-08 01:05:38 +00001202<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001203
Chris Lattner261efe92003-11-25 01:02:51 +00001204<p>The LLVM instruction set consists of several different
1205classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +00001206instructions</a>, <a href="#binaryops">binary instructions</a>,
1207<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +00001208 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1209instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001210
Misha Brukman9d0919f2003-11-08 01:05:38 +00001211</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001212
Chris Lattner00950542001-06-06 20:29:01 +00001213<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001214<div class="doc_subsection"> <a name="terminators">Terminator
1215Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001216
Misha Brukman9d0919f2003-11-08 01:05:38 +00001217<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001218
Chris Lattner261efe92003-11-25 01:02:51 +00001219<p>As mentioned <a href="#functionstructure">previously</a>, every
1220basic block in a program ends with a "Terminator" instruction, which
1221indicates which block should be executed after the current block is
1222finished. These terminator instructions typically yield a '<tt>void</tt>'
1223value: they produce control flow, not values (the one exception being
1224the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +00001225<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +00001226 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1227instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +00001228the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1229 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1230 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001231
Misha Brukman9d0919f2003-11-08 01:05:38 +00001232</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001233
Chris Lattner00950542001-06-06 20:29:01 +00001234<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001235<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1236Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001237<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001238<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001239<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 +00001240 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001241</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001242<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001243<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +00001244value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +00001245<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +00001246returns a value and then causes control flow, and one that just causes
1247control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001248<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001249<p>The '<tt>ret</tt>' instruction may return any '<a
1250 href="#t_firstclass">first class</a>' type. Notice that a function is
1251not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1252instruction inside of the function that returns a value that does not
1253match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001254<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001255<p>When the '<tt>ret</tt>' instruction is executed, control flow
1256returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +00001257 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +00001258the instruction after the call. If the caller was an "<a
1259 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +00001260at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +00001261returns a value, that value shall set the call or invoke instruction's
1262return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001263<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001264<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001265 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001266</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001267</div>
Chris Lattner00950542001-06-06 20:29:01 +00001268<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001269<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001270<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001271<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001272<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 +00001273</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001274<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001275<p>The '<tt>br</tt>' instruction is used to cause control flow to
1276transfer to a different basic block in the current function. There are
1277two forms of this instruction, corresponding to a conditional branch
1278and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001279<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001280<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1281single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1282unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1283value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001284<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001285<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1286argument is evaluated. If the value is <tt>true</tt>, control flows
1287to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1288control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001289<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001290<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
1291 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 +00001292</div>
Chris Lattner00950542001-06-06 20:29:01 +00001293<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001294<div class="doc_subsubsection">
1295 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1296</div>
1297
Misha Brukman9d0919f2003-11-08 01:05:38 +00001298<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001299<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001300
1301<pre>
1302 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1303</pre>
1304
Chris Lattner00950542001-06-06 20:29:01 +00001305<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001306
1307<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1308several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +00001309instruction, allowing a branch to occur to one of many possible
1310destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001311
1312
Chris Lattner00950542001-06-06 20:29:01 +00001313<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001314
1315<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1316comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1317an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1318table is not allowed to contain duplicate constant entries.</p>
1319
Chris Lattner00950542001-06-06 20:29:01 +00001320<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001321
Chris Lattner261efe92003-11-25 01:02:51 +00001322<p>The <tt>switch</tt> instruction specifies a table of values and
1323destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +00001324table is searched for the given value. If the value is found, control flow is
1325transfered to the corresponding destination; otherwise, control flow is
1326transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001327
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001328<h5>Implementation:</h5>
1329
1330<p>Depending on properties of the target machine and the particular
1331<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +00001332ways. For example, it could be generated as a series of chained conditional
1333branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001334
1335<h5>Example:</h5>
1336
1337<pre>
1338 <i>; Emulate a conditional br instruction</i>
1339 %Val = <a href="#i_cast">cast</a> bool %value to int
1340 switch int %Val, label %truedest [int 0, label %falsedest ]
1341
1342 <i>; Emulate an unconditional br instruction</i>
1343 switch uint 0, label %dest [ ]
1344
1345 <i>; Implement a jump table:</i>
1346 switch uint %val, label %otherwise [ uint 0, label %onzero
1347 uint 1, label %onone
1348 uint 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00001349</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001350</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001351
Chris Lattner00950542001-06-06 20:29:01 +00001352<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001353<div class="doc_subsubsection">
1354 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1355</div>
1356
Misha Brukman9d0919f2003-11-08 01:05:38 +00001357<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001358
Chris Lattner00950542001-06-06 20:29:01 +00001359<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001360
1361<pre>
1362 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)
1363 to label &lt;normal label&gt; except label &lt;exception label&gt;
1364</pre>
1365
Chris Lattner6536cfe2002-05-06 22:08:29 +00001366<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001367
1368<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1369function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +00001370'<tt>normal</tt>' label or the
1371'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001372"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1373"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +00001374href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1375continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001376
Chris Lattner00950542001-06-06 20:29:01 +00001377<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001378
Misha Brukman9d0919f2003-11-08 01:05:38 +00001379<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001380
Chris Lattner00950542001-06-06 20:29:01 +00001381<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001382 <li>
John Criswellc1f786c2005-05-13 22:25:59 +00001383 The optional "cconv" marker indicates which <a href="callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001384 convention</a> the call should use. If none is specified, the call defaults
1385 to using C calling conventions.
1386 </li>
1387 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1388 function value being invoked. In most cases, this is a direct function
1389 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1390 an arbitrary pointer to function value.
1391 </li>
1392
1393 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1394 function to be invoked. </li>
1395
1396 <li>'<tt>function args</tt>': argument list whose types match the function
1397 signature argument types. If the function signature indicates the function
1398 accepts a variable number of arguments, the extra arguments can be
1399 specified. </li>
1400
1401 <li>'<tt>normal label</tt>': the label reached when the called function
1402 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1403
1404 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1405 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1406
Chris Lattner00950542001-06-06 20:29:01 +00001407</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001408
Chris Lattner00950542001-06-06 20:29:01 +00001409<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001410
Misha Brukman9d0919f2003-11-08 01:05:38 +00001411<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001412href="#i_call">call</a></tt>' instruction in most regards. The primary
1413difference is that it establishes an association with a label, which is used by
1414the runtime library to unwind the stack.</p>
1415
1416<p>This instruction is used in languages with destructors to ensure that proper
1417cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1418exception. Additionally, this is important for implementation of
1419'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1420
Chris Lattner00950542001-06-06 20:29:01 +00001421<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001422<pre>
1423 %retval = invoke int %Test(int 15) to label %Continue
1424 except label %TestCleanup <i>; {int}:retval set</i>
1425 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
1426 except label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00001427</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001428</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001429
1430
Chris Lattner27f71f22003-09-03 00:41:47 +00001431<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00001432
Chris Lattner261efe92003-11-25 01:02:51 +00001433<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1434Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +00001435
Misha Brukman9d0919f2003-11-08 01:05:38 +00001436<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +00001437
Chris Lattner27f71f22003-09-03 00:41:47 +00001438<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001439<pre>
1440 unwind
1441</pre>
1442
Chris Lattner27f71f22003-09-03 00:41:47 +00001443<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001444
1445<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1446at the first callee in the dynamic call stack which used an <a
1447href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1448primarily used to implement exception handling.</p>
1449
Chris Lattner27f71f22003-09-03 00:41:47 +00001450<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001451
1452<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1453immediately halt. The dynamic call stack is then searched for the first <a
1454href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1455execution continues at the "exceptional" destination block specified by the
1456<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1457dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001458</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001459
1460<!-- _______________________________________________________________________ -->
1461
1462<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1463Instruction</a> </div>
1464
1465<div class="doc_text">
1466
1467<h5>Syntax:</h5>
1468<pre>
1469 unreachable
1470</pre>
1471
1472<h5>Overview:</h5>
1473
1474<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1475instruction is used to inform the optimizer that a particular portion of the
1476code is not reachable. This can be used to indicate that the code after a
1477no-return function cannot be reached, and other facts.</p>
1478
1479<h5>Semantics:</h5>
1480
1481<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1482</div>
1483
1484
1485
Chris Lattner00950542001-06-06 20:29:01 +00001486<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001487<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001488<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001489<p>Binary operators are used to do most of the computation in a
1490program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +00001491produce a single value. The operands might represent
Chris Lattnera58561b2004-08-12 19:12:28 +00001492multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1493The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +00001494necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001495<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001496</div>
Chris Lattner00950542001-06-06 20:29:01 +00001497<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001498<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1499Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001500<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001501<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001502<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 +00001503</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001504<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001505<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001506<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001507<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +00001508 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1509 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1510Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001511<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001512<p>The value produced is the integer or floating point sum of the two
1513operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001514<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001515<pre> &lt;result&gt; = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001516</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001517</div>
Chris Lattner00950542001-06-06 20:29:01 +00001518<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001519<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1520Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001521<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001522<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001523<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 +00001524</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001525<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001526<p>The '<tt>sub</tt>' instruction returns the difference of its two
1527operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001528<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1529instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001530<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001531<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001532 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001533values.
1534This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1535Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001536<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001537<p>The value produced is the integer or floating point difference of
1538the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001539<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001540<pre> &lt;result&gt; = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001541 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1542</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001543</div>
Chris Lattner00950542001-06-06 20:29:01 +00001544<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001545<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1546Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001547<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001548<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001549<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 +00001550</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001551<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001552<p>The '<tt>mul</tt>' instruction returns the product of its two
1553operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001554<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001555<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001556 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001557values.
1558This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1559Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001560<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001561<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +00001562two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001563<p>There is no signed vs unsigned multiplication. The appropriate
1564action is taken based on the type of the operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001565<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001566<pre> &lt;result&gt; = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001567</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001568</div>
Chris Lattner00950542001-06-06 20:29:01 +00001569<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001570<div class="doc_subsubsection"> <a name="i_div">'<tt>div</tt>'
1571Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001572<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001573<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001574<pre> &lt;result&gt; = div &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1575</pre>
1576<h5>Overview:</h5>
1577<p>The '<tt>div</tt>' instruction returns the quotient of its two
1578operands.</p>
1579<h5>Arguments:</h5>
1580<p>The two arguments to the '<tt>div</tt>' instruction must be either <a
1581 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001582values.
1583This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1584Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001585<h5>Semantics:</h5>
1586<p>The value produced is the integer or floating point quotient of the
1587two operands.</p>
1588<h5>Example:</h5>
1589<pre> &lt;result&gt; = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1590</pre>
1591</div>
1592<!-- _______________________________________________________________________ -->
1593<div class="doc_subsubsection"> <a name="i_rem">'<tt>rem</tt>'
1594Instruction</a> </div>
1595<div class="doc_text">
1596<h5>Syntax:</h5>
1597<pre> &lt;result&gt; = rem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1598</pre>
1599<h5>Overview:</h5>
1600<p>The '<tt>rem</tt>' instruction returns the remainder from the
1601division of its two operands.</p>
1602<h5>Arguments:</h5>
1603<p>The two arguments to the '<tt>rem</tt>' instruction must be either <a
1604 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001605values.
1606This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1607Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001608<h5>Semantics:</h5>
1609<p>This returns the <i>remainder</i> of a division (where the result
1610has the same sign as the divisor), not the <i>modulus</i> (where the
1611result has the same sign as the dividend) of a value. For more
John Criswell0ec250c2005-10-24 16:17:18 +00001612information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +00001613 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1614Math Forum</a>.</p>
1615<h5>Example:</h5>
1616<pre> &lt;result&gt; = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1617</pre>
Robert Bocchino7b81c752006-02-17 21:18:08 +00001618
Chris Lattner261efe92003-11-25 01:02:51 +00001619</div>
1620<!-- _______________________________________________________________________ -->
1621<div class="doc_subsubsection"> <a name="i_setcc">'<tt>set<i>cc</i></tt>'
1622Instructions</a> </div>
1623<div class="doc_text">
1624<h5>Syntax:</h5>
1625<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 +00001626 &lt;result&gt; = setne &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1627 &lt;result&gt; = setlt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1628 &lt;result&gt; = setgt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1629 &lt;result&gt; = setle &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1630 &lt;result&gt; = setge &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1631</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001632<h5>Overview:</h5>
1633<p>The '<tt>set<i>cc</i></tt>' family of instructions returns a boolean
1634value based on a comparison of their two operands.</p>
1635<h5>Arguments:</h5>
1636<p>The two arguments to the '<tt>set<i>cc</i></tt>' instructions must
1637be of <a href="#t_firstclass">first class</a> type (it is not possible
1638to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1639or '<tt>void</tt>' values, etc...). Both arguments must have identical
1640types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001641<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001642<p>The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1643value if both operands are equal.<br>
1644The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1645value if both operands are unequal.<br>
1646The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1647value if the first operand is less than the second operand.<br>
1648The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1649value if the first operand is greater than the second operand.<br>
1650The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1651value if the first operand is less than or equal to the second operand.<br>
1652The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1653value if the first operand is greater than or equal to the second
1654operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001655<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001656<pre> &lt;result&gt; = seteq int 4, 5 <i>; yields {bool}:result = false</i>
Chris Lattner00950542001-06-06 20:29:01 +00001657 &lt;result&gt; = setne float 4, 5 <i>; yields {bool}:result = true</i>
1658 &lt;result&gt; = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1659 &lt;result&gt; = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1660 &lt;result&gt; = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1661 &lt;result&gt; = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1662</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001663</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +00001664
Chris Lattner00950542001-06-06 20:29:01 +00001665<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001666<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1667Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001668<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001669<p>Bitwise binary operators are used to do various forms of
1670bit-twiddling in a program. They are generally very efficient
John Criswell9e2485c2004-12-10 15:51:16 +00001671instructions and can commonly be strength reduced from other
Chris Lattner261efe92003-11-25 01:02:51 +00001672instructions. They require two operands, execute an operation on them,
1673and produce a single value. The resulting value of the bitwise binary
1674operators is always the same type as its first operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001675</div>
Chris Lattner00950542001-06-06 20:29:01 +00001676<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001677<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1678Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001679<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001680<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001681<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 +00001682</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001683<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001684<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1685its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001686<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001687<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001688 href="#t_integral">integral</a> values. Both arguments must have
1689identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001690<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001691<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001692<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001693<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +00001694<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001695 <tbody>
1696 <tr>
1697 <td>In0</td>
1698 <td>In1</td>
1699 <td>Out</td>
1700 </tr>
1701 <tr>
1702 <td>0</td>
1703 <td>0</td>
1704 <td>0</td>
1705 </tr>
1706 <tr>
1707 <td>0</td>
1708 <td>1</td>
1709 <td>0</td>
1710 </tr>
1711 <tr>
1712 <td>1</td>
1713 <td>0</td>
1714 <td>0</td>
1715 </tr>
1716 <tr>
1717 <td>1</td>
1718 <td>1</td>
1719 <td>1</td>
1720 </tr>
1721 </tbody>
1722</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001723</div>
Chris Lattner00950542001-06-06 20:29:01 +00001724<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001725<pre> &lt;result&gt; = and int 4, %var <i>; yields {int}:result = 4 &amp; %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001726 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1727 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1728</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001729</div>
Chris Lattner00950542001-06-06 20:29:01 +00001730<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001731<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001732<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001733<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001734<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 +00001735</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001736<h5>Overview:</h5>
1737<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1738or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001739<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001740<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001741 href="#t_integral">integral</a> values. Both arguments must have
1742identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001743<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001744<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001745<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001746<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001747<table border="1" cellspacing="0" cellpadding="4">
1748 <tbody>
1749 <tr>
1750 <td>In0</td>
1751 <td>In1</td>
1752 <td>Out</td>
1753 </tr>
1754 <tr>
1755 <td>0</td>
1756 <td>0</td>
1757 <td>0</td>
1758 </tr>
1759 <tr>
1760 <td>0</td>
1761 <td>1</td>
1762 <td>1</td>
1763 </tr>
1764 <tr>
1765 <td>1</td>
1766 <td>0</td>
1767 <td>1</td>
1768 </tr>
1769 <tr>
1770 <td>1</td>
1771 <td>1</td>
1772 <td>1</td>
1773 </tr>
1774 </tbody>
1775</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001776</div>
Chris Lattner00950542001-06-06 20:29:01 +00001777<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001778<pre> &lt;result&gt; = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001779 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1780 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1781</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001782</div>
Chris Lattner00950542001-06-06 20:29:01 +00001783<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001784<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1785Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001786<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001787<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001788<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 +00001789</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001790<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001791<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1792or of its two operands. The <tt>xor</tt> is used to implement the
1793"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001794<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001795<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001796 href="#t_integral">integral</a> values. Both arguments must have
1797identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001798<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001799<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001800<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001801<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001802<table border="1" cellspacing="0" cellpadding="4">
1803 <tbody>
1804 <tr>
1805 <td>In0</td>
1806 <td>In1</td>
1807 <td>Out</td>
1808 </tr>
1809 <tr>
1810 <td>0</td>
1811 <td>0</td>
1812 <td>0</td>
1813 </tr>
1814 <tr>
1815 <td>0</td>
1816 <td>1</td>
1817 <td>1</td>
1818 </tr>
1819 <tr>
1820 <td>1</td>
1821 <td>0</td>
1822 <td>1</td>
1823 </tr>
1824 <tr>
1825 <td>1</td>
1826 <td>1</td>
1827 <td>0</td>
1828 </tr>
1829 </tbody>
1830</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001831</div>
Chris Lattner261efe92003-11-25 01:02:51 +00001832<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +00001833<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001834<pre> &lt;result&gt; = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001835 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
1836 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner27f71f22003-09-03 00:41:47 +00001837 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00001838</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001839</div>
Chris Lattner00950542001-06-06 20:29:01 +00001840<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001841<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
1842Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001843<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001844<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001845<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 +00001846</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001847<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001848<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
1849the left a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001850<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001851<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001852 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1853type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001854<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001855<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001856<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001857<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 +00001858 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1859 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1860</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001861</div>
Chris Lattner00950542001-06-06 20:29:01 +00001862<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001863<div class="doc_subsubsection"> <a name="i_shr">'<tt>shr</tt>'
1864Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001865<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001866<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001867<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 +00001868</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001869<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001870<p>The '<tt>shr</tt>' instruction returns the first operand shifted to
1871the right a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001872<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001873<p>The first argument to the '<tt>shr</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001874 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1875type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001876<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001877<p>If the first argument is a <a href="#t_signed">signed</a> type, the
1878most significant bit is duplicated in the newly free'd bit positions.
1879If the first argument is unsigned, zero bits shall fill the empty
1880positions.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001881<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001882<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 +00001883 &lt;result&gt; = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
Chris Lattner00950542001-06-06 20:29:01 +00001884 &lt;result&gt; = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +00001885 &lt;result&gt; = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1886 &lt;result&gt; = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner00950542001-06-06 20:29:01 +00001887</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001888</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001889
Chris Lattner00950542001-06-06 20:29:01 +00001890<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001891<div class="doc_subsection">
1892 <a name="memoryops">Memory Access Operations</a>
1893</div>
1894
Misha Brukman9d0919f2003-11-08 01:05:38 +00001895<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001896
Chris Lattner261efe92003-11-25 01:02:51 +00001897<p>A key design point of an SSA-based representation is how it
1898represents memory. In LLVM, no memory locations are in SSA form, which
1899makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +00001900allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001901
Misha Brukman9d0919f2003-11-08 01:05:38 +00001902</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001903
Chris Lattner00950542001-06-06 20:29:01 +00001904<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001905<div class="doc_subsubsection">
1906 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
1907</div>
1908
Misha Brukman9d0919f2003-11-08 01:05:38 +00001909<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001910
Chris Lattner00950542001-06-06 20:29:01 +00001911<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001912
1913<pre>
1914 &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 +00001915</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001916
Chris Lattner00950542001-06-06 20:29:01 +00001917<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001918
Chris Lattner261efe92003-11-25 01:02:51 +00001919<p>The '<tt>malloc</tt>' instruction allocates memory from the system
1920heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001921
Chris Lattner00950542001-06-06 20:29:01 +00001922<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001923
1924<p>The '<tt>malloc</tt>' instruction allocates
1925<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +00001926bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00001927appropriate type to the program. If "NumElements" is specified, it is the
1928number of elements allocated. If an alignment is specified, the value result
1929of the allocation is guaranteed to be aligned to at least that boundary. If
1930not specified, or if zero, the target can choose to align the allocation on any
1931convenient boundary.</p>
1932
Misha Brukman9d0919f2003-11-08 01:05:38 +00001933<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001934
Chris Lattner00950542001-06-06 20:29:01 +00001935<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001936
Chris Lattner261efe92003-11-25 01:02:51 +00001937<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
1938a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001939
Chris Lattner2cbdc452005-11-06 08:02:57 +00001940<h5>Example:</h5>
1941
1942<pre>
1943 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
1944
1945 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001946 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1947 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001948 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
1949 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +00001950</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001951</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001952
Chris Lattner00950542001-06-06 20:29:01 +00001953<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001954<div class="doc_subsubsection">
1955 <a name="i_free">'<tt>free</tt>' Instruction</a>
1956</div>
1957
Misha Brukman9d0919f2003-11-08 01:05:38 +00001958<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001959
Chris Lattner00950542001-06-06 20:29:01 +00001960<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001961
1962<pre>
1963 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +00001964</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001965
Chris Lattner00950542001-06-06 20:29:01 +00001966<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001967
Chris Lattner261efe92003-11-25 01:02:51 +00001968<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +00001969memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001970
Chris Lattner00950542001-06-06 20:29:01 +00001971<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001972
Chris Lattner261efe92003-11-25 01:02:51 +00001973<p>'<tt>value</tt>' shall be a pointer value that points to a value
1974that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
1975instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001976
Chris Lattner00950542001-06-06 20:29:01 +00001977<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001978
John Criswell9e2485c2004-12-10 15:51:16 +00001979<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +00001980after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001981
Chris Lattner00950542001-06-06 20:29:01 +00001982<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001983
1984<pre>
1985 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
Chris Lattner00950542001-06-06 20:29:01 +00001986 free [4 x ubyte]* %array
1987</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001988</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001989
Chris Lattner00950542001-06-06 20:29:01 +00001990<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001991<div class="doc_subsubsection">
1992 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
1993</div>
1994
Misha Brukman9d0919f2003-11-08 01:05:38 +00001995<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00001996
Chris Lattner00950542001-06-06 20:29:01 +00001997<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001998
1999<pre>
2000 &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 +00002001</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002002
Chris Lattner00950542001-06-06 20:29:01 +00002003<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002004
Chris Lattner261efe92003-11-25 01:02:51 +00002005<p>The '<tt>alloca</tt>' instruction allocates memory on the current
2006stack frame of the procedure that is live until the current function
2007returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002008
Chris Lattner00950542001-06-06 20:29:01 +00002009<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002010
John Criswell9e2485c2004-12-10 15:51:16 +00002011<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002012bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00002013appropriate type to the program. If "NumElements" is specified, it is the
2014number of elements allocated. If an alignment is specified, the value result
2015of the allocation is guaranteed to be aligned to at least that boundary. If
2016not specified, or if zero, the target can choose to align the allocation on any
2017convenient boundary.</p>
2018
Misha Brukman9d0919f2003-11-08 01:05:38 +00002019<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002020
Chris Lattner00950542001-06-06 20:29:01 +00002021<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002022
John Criswellc1f786c2005-05-13 22:25:59 +00002023<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +00002024memory is automatically released when the function returns. The '<tt>alloca</tt>'
2025instruction is commonly used to represent automatic variables that must
2026have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +00002027 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002028instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002029
Chris Lattner00950542001-06-06 20:29:01 +00002030<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002031
2032<pre>
2033 %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002034 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002035 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
2036 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00002037</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002038</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002039
Chris Lattner00950542001-06-06 20:29:01 +00002040<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002041<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2042Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002043<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00002044<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002045<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 +00002046<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002047<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002048<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002049<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +00002050address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +00002051 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +00002052marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +00002053the number or order of execution of this <tt>load</tt> with other
2054volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2055instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002056<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002057<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002058<h5>Examples:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002059<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2060 <a
2061 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002062 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2063</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002064</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002065<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002066<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2067Instruction</a> </div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002068<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002069<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 +00002070 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 +00002071</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002072<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002073<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002074<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002075<p>There are two arguments to the '<tt>store</tt>' instruction: a value
John Criswell0ec250c2005-10-24 16:17:18 +00002076to 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 +00002077operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +00002078operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +00002079optimizer is not allowed to modify the number or order of execution of
2080this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2081 href="#i_store">store</a></tt> instructions.</p>
2082<h5>Semantics:</h5>
2083<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2084at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002085<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002086<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2087 <a
2088 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002089 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2090</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002091<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002092<div class="doc_subsubsection">
2093 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2094</div>
2095
Misha Brukman9d0919f2003-11-08 01:05:38 +00002096<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +00002097<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002098<pre>
2099 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2100</pre>
2101
Chris Lattner7faa8832002-04-14 06:13:44 +00002102<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002103
2104<p>
2105The '<tt>getelementptr</tt>' instruction is used to get the address of a
2106subelement of an aggregate data structure.</p>
2107
Chris Lattner7faa8832002-04-14 06:13:44 +00002108<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002109
2110<p>This instruction takes a list of integer constants that indicate what
2111elements of the aggregate object to index to. The actual types of the arguments
2112provided depend on the type of the first pointer argument. The
2113'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +00002114levels of a structure or to a specific index in an array. When indexing into a
2115structure, only <tt>uint</tt>
John Criswellc1f786c2005-05-13 22:25:59 +00002116integer constants are allowed. When indexing into an array or pointer,
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002117<tt>int</tt> and <tt>long</tt> indexes are allowed of any sign.</p>
2118
Chris Lattner261efe92003-11-25 01:02:51 +00002119<p>For example, let's consider a C code fragment and how it gets
2120compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002121
2122<pre>
2123 struct RT {
2124 char A;
2125 int B[10][20];
2126 char C;
2127 };
2128 struct ST {
2129 int X;
2130 double Y;
2131 struct RT Z;
2132 };
2133
2134 int *foo(struct ST *s) {
2135 return &amp;s[1].Z.B[5][13];
2136 }
2137</pre>
2138
Misha Brukman9d0919f2003-11-08 01:05:38 +00002139<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002140
2141<pre>
2142 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2143 %ST = type { int, double, %RT }
2144
Brian Gaeke7283e7c2004-07-02 21:08:14 +00002145 implementation
2146
2147 int* %foo(%ST* %s) {
2148 entry:
2149 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002150 ret int* %reg
2151 }
2152</pre>
2153
Chris Lattner7faa8832002-04-14 06:13:44 +00002154<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002155
2156<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +00002157on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Chris Lattnere53e5082004-06-03 22:57:15 +00002158and <a href="#t_array">array</a> types require <tt>uint</tt>, <tt>int</tt>,
2159<tt>ulong</tt>, or <tt>long</tt> values, and <a href="#t_struct">structure</a>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002160types require <tt>uint</tt> <b>constants</b>.</p>
2161
Misha Brukman9d0919f2003-11-08 01:05:38 +00002162<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002163type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2164}</tt>' type, a structure. The second index indexes into the third element of
2165the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2166sbyte }</tt>' type, another structure. The third index indexes into the second
2167element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2168array. The two dimensions of the array are subscripted into, yielding an
John Criswellfc6b8952005-05-16 16:17:45 +00002169'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002170to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2171
Chris Lattner261efe92003-11-25 01:02:51 +00002172<p>Note that it is perfectly legal to index partially through a
2173structure, returning a pointer to an inner element. Because of this,
2174the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002175
2176<pre>
Chris Lattnerd4f6b172005-03-07 22:13:59 +00002177 int* %foo(%ST* %s) {
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002178 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2179 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2180 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2181 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2182 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2183 ret int* %t5
2184 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00002185</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00002186
2187<p>Note that it is undefined to access an array out of bounds: array and
2188pointer indexes must always be within the defined bounds of the array type.
2189The one exception for this rules is zero length arrays. These arrays are
2190defined to be accessible as variable length arrays, which requires access
2191beyond the zero'th element.</p>
2192
Chris Lattner7faa8832002-04-14 06:13:44 +00002193<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +00002194
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002195<pre>
2196 <i>; yields [12 x ubyte]*:aptr</i>
2197 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2198</pre>
2199
2200</div>
Chris Lattner00950542001-06-06 20:29:01 +00002201<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00002202<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002203<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +00002204<p>The instructions in this category are the "miscellaneous"
Chris Lattner261efe92003-11-25 01:02:51 +00002205instructions, which defy better classification.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002206</div>
Chris Lattner00950542001-06-06 20:29:01 +00002207<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002208<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
2209Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002210<div class="doc_text">
Chris Lattner33ba0d92001-07-09 00:26:23 +00002211<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002212<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002213<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002214<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
2215the SSA graph representing the function.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002216<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002217<p>The type of the incoming values are specified with the first type
2218field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
2219as arguments, with one pair for each predecessor basic block of the
2220current block. Only values of <a href="#t_firstclass">first class</a>
2221type may be used as the value arguments to the PHI node. Only labels
2222may be used as the label arguments.</p>
2223<p>There must be no non-phi instructions between the start of a basic
2224block and the PHI instructions: i.e. PHI instructions must be first in
2225a basic block.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002226<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002227<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
2228value specified by the parameter, depending on which basic block we
2229came from in the last <a href="#terminators">terminator</a> instruction.</p>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002230<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002231<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 +00002232</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002233
Chris Lattner6536cfe2002-05-06 22:08:29 +00002234<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +00002235<div class="doc_subsubsection">
2236 <a name="i_cast">'<tt>cast .. to</tt>' Instruction</a>
2237</div>
2238
Misha Brukman9d0919f2003-11-08 01:05:38 +00002239<div class="doc_text">
Chris Lattnercc37aae2004-03-12 05:50:16 +00002240
Chris Lattner6536cfe2002-05-06 22:08:29 +00002241<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002242
2243<pre>
2244 &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 +00002245</pre>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002246
Chris Lattner6536cfe2002-05-06 22:08:29 +00002247<h5>Overview:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002248
2249<p>
2250The '<tt>cast</tt>' instruction is used as the primitive means to convert
2251integers to floating point, change data type sizes, and break type safety (by
2252casting pointers).
2253</p>
2254
2255
Chris Lattner6536cfe2002-05-06 22:08:29 +00002256<h5>Arguments:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002257
2258<p>
2259The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
2260class value, and a type to cast it to, which must also be a <a
2261href="#t_firstclass">first class</a> type.
2262</p>
2263
Chris Lattner6536cfe2002-05-06 22:08:29 +00002264<h5>Semantics:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002265
2266<p>
2267This instruction follows the C rules for explicit casts when determining how the
2268data being cast must change to fit in its new container.
2269</p>
2270
2271<p>
2272When casting to bool, any value that would be considered true in the context of
2273a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
2274all else are '<tt>false</tt>'.
2275</p>
2276
2277<p>
2278When extending an integral value from a type of one signness to another (for
2279example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
2280<b>source</b> value is signed, and zero-extended if the source value is
2281unsigned. <tt>bool</tt> values are always zero extended into either zero or
2282one.
2283</p>
2284
Chris Lattner33ba0d92001-07-09 00:26:23 +00002285<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002286
2287<pre>
2288 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
Chris Lattner7bae3952002-06-25 18:03:17 +00002289 %Y = cast int 123 to bool <i>; yields bool:true</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002290</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002291</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002292
2293<!-- _______________________________________________________________________ -->
2294<div class="doc_subsubsection">
2295 <a name="i_select">'<tt>select</tt>' Instruction</a>
2296</div>
2297
2298<div class="doc_text">
2299
2300<h5>Syntax:</h5>
2301
2302<pre>
2303 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
2304</pre>
2305
2306<h5>Overview:</h5>
2307
2308<p>
2309The '<tt>select</tt>' instruction is used to choose one value based on a
2310condition, without branching.
2311</p>
2312
2313
2314<h5>Arguments:</h5>
2315
2316<p>
2317The '<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.
2318</p>
2319
2320<h5>Semantics:</h5>
2321
2322<p>
2323If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +00002324value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +00002325</p>
2326
2327<h5>Example:</h5>
2328
2329<pre>
2330 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
2331</pre>
2332</div>
2333
Robert Bocchino7b81c752006-02-17 21:18:08 +00002334<!-- _______________________________________________________________________ -->
Robert Bocchino8949de22006-03-14 20:55:28 +00002335<div class="doc_subsubsection"> <a name="i_vsetint">'<tt>vsetint</tt>'
Robert Bocchino7b81c752006-02-17 21:18:08 +00002336Instruction</a> </div>
2337<div class="doc_text">
2338<h5>Syntax:</h5>
Robert Bocchino8949de22006-03-14 20:55:28 +00002339<pre>&lt;result&gt; = vsetint &lt;op&gt;, &lt;n x &lt;ty&gt;&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields &lt;n x bool&gt;</i>
Robert Bocchino7b81c752006-02-17 21:18:08 +00002340</pre>
2341
2342<h5>Overview:</h5>
2343
Robert Bocchino8949de22006-03-14 20:55:28 +00002344<p>The '<tt>vsetint</tt>' instruction takes two integer vectors and
2345returns a vector of boolean values representing, at each position, the
2346result of the comparison between the values at that position in the
2347two operands.</p>
Robert Bocchino7b81c752006-02-17 21:18:08 +00002348
2349<h5>Arguments:</h5>
2350
Robert Bocchino8949de22006-03-14 20:55:28 +00002351<p>The arguments to a '<tt>vsetint</tt>' instruction are a comparison
Robert Bocchino7b81c752006-02-17 21:18:08 +00002352operation and two value arguments. The value arguments must be of <a
Robert Bocchino8949de22006-03-14 20:55:28 +00002353href="#t_integral">integral</a> <a href="#t_packed">packed</a> type,
2354and they must have identical types. The operation argument must be
2355one of <tt>eq</tt>, <tt>ne</tt>, <tt>slt</tt>, <tt>sgt</tt>,
2356<tt>sle</tt>, <tt>sge</tt>, <tt>ult</tt>, <tt>ugt</tt>, <tt>ule</tt>,
2357<tt>uge</tt>, <tt>true</tt>, and <tt>false</tt>. The result is a
2358packed <tt>bool</tt> value with the same length as each operand.</p>
Robert Bocchino7b81c752006-02-17 21:18:08 +00002359
2360<h5>Semantics:</h5>
2361
Robert Bocchino8949de22006-03-14 20:55:28 +00002362<p>The following table shows the semantics of '<tt>vsetint</tt>'. For
2363each position of the result, the comparison is done on the
2364corresponding positions of the two value arguments. Note that the
2365signedness of the comparison depends on the comparison opcode and
2366<i>not</i> on the signedness of the value operands. E.g., <tt>vsetint
2367slt <4 x unsigned> %x, %y</tt> does an elementwise <i>signed</i>
2368comparison of <tt>%x</tt> and <tt>%y</tt>.</p>
Robert Bocchino7b81c752006-02-17 21:18:08 +00002369
2370<table border="1" cellspacing="0" cellpadding="4">
2371 <tbody>
2372 <tr><th>Operation</th><th>Result is true iff</th><th>Comparison is</th></tr>
2373 <tr><td><tt>eq</tt></td><td>var1 == var2</td><td>--</td></tr>
2374 <tr><td><tt>ne</tt></td><td>var1 != var2</td><td>--</td></tr>
Robert Bocchino8949de22006-03-14 20:55:28 +00002375 <tr><td><tt>slt</tt></td><td>var1 &lt; var2</td><td>signed</td></tr>
2376 <tr><td><tt>sgt</tt></td><td>var1 &gt; var2</td><td>signed</td></tr>
2377 <tr><td><tt>sle</tt></td><td>var1 &lt;= var2</td><td>signed</td></tr>
2378 <tr><td><tt>sge</tt></td><td>var1 &gt;= var2</td><td>signed</td></tr>
Robert Bocchino7b81c752006-02-17 21:18:08 +00002379 <tr><td><tt>ult</tt></td><td>var1 &lt; var2</td><td>unsigned</td></tr>
2380 <tr><td><tt>ugt</tt></td><td>var1 &gt; var2</td><td>unsigned</td></tr>
2381 <tr><td><tt>ule</tt></td><td>var1 &lt;= var2</td><td>unsigned</td></tr>
2382 <tr><td><tt>uge</tt></td><td>var1 &gt;= var2</td><td>unsigned</td></tr>
2383 <tr><td><tt>true</tt></td><td>always</td><td>--</td></tr>
2384 <tr><td><tt>false</tt></td><td>never</td><td>--</td></tr>
2385 </tbody>
2386</table>
2387
Robert Bocchino8949de22006-03-14 20:55:28 +00002388<h5>Example:</h5>
2389<pre> &lt;result&gt; = vsetint eq &lt;2 x int&gt; &lt;int 0, int 1&gt;, &lt;int 1, int 0&gt; <i>; yields {&lt;2 x bool&gt;}:result = false, false</i>
2390 &lt;result&gt; = vsetint ne &lt;2 x int&gt; &lt;int 0, int 1&gt;, &lt;int 1, int 0&gt; <i>; yields {&lt;2 x bool&gt;}:result = true, true</i>
2391 &lt;result&gt; = vsetint slt &lt;2 x int&gt; &lt;int 0, int 1&gt;, &lt;int 1, int 0&gt; <i>; yields {&lt;2 x bool&gt;}:result = true, false</i>
2392 &lt;result&gt; = vsetint sgt &lt;2 x int&gt; &lt;int 0, int 1&gt;, &lt;int 1, int 0&gt; <i>; yields {&lt;2 x bool&gt;}:result = false, true</i>
2393 &lt;result&gt; = vsetint sle &lt;2 x int&gt; &lt;int 0, int 1&gt;, &lt;int 1, int 0&gt; <i>; yields {&lt;2 x bool&gt;}:result = true, false</i>
2394 &lt;result&gt; = vsetint sge &lt;2 x int&gt; &lt;int 0, int 1&gt;, &lt;int 1, int 0&gt; <i>; yields {&lt;2 x bool&gt;}:result = false, true</i>
2395</pre>
2396</div>
2397
2398<!-- _______________________________________________________________________ -->
2399<div class="doc_subsubsection"> <a name="i_vsetfp">'<tt>vsetfp</tt>'
2400Instruction</a> </div>
2401<div class="doc_text">
2402<h5>Syntax:</h5>
2403<pre>&lt;result&gt; = vsetfp &lt;op&gt;, &lt;n x &lt;ty&gt;&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields &lt;n x bool&gt;</i>
2404</pre>
2405
2406<h5>Overview:</h5>
2407
2408<p>The '<tt>vsetfp</tt>' instruction takes two floating point vector
2409arguments and returns a vector of boolean values representing, at each
2410position, the result of the comparison between the values at that
2411position in the two operands.</p>
2412
2413<h5>Arguments:</h5>
2414
2415<p>The arguments to a '<tt>vsetfp</tt>' instruction are a comparison
2416operation and two value arguments. The value arguments must be of <a
2417href="t_floating">floating point</a> <a href="#t_packed">packed</a>
2418type, and they must have identical types. The operation argument must
2419be one of <tt>eq</tt>, <tt>ne</tt>, <tt>lt</tt>, <tt>gt</tt>,
2420<tt>le</tt>, <tt>ge</tt>, <tt>oeq</tt>, <tt>one</tt>, <tt>olt</tt>,
2421<tt>ogt</tt>, <tt>ole</tt>, <tt>oge</tt>, <tt>ueq</tt>, <tt>une</tt>,
2422<tt>ult</tt>, <tt>ugt</tt>, <tt>ule</tt>, <tt>uge</tt>, <tt>o</tt>,
2423<tt>u</tt>, <tt>true</tt>, and <tt>false</tt>. The result is a packed
2424<tt>bool</tt> value with the same length as each operand.</p>
2425
2426<h5>Semantics:</h5>
2427
2428<p>The following table shows the semantics of '<tt>vsetfp</tt>' for
Robert Bocchino7b81c752006-02-17 21:18:08 +00002429floating point types. If either operand is a floating point Not a
2430Number (NaN) value, the operation is unordered, and the value in the
2431first column below is produced at that position. Otherwise, the
2432operation is ordered, and the value in the second column is
2433produced.</p>
2434
2435<table border="1" cellspacing="0" cellpadding="4">
2436 <tbody>
2437 <tr><th>Operation</th><th>If unordered<th>Otherwise true iff</th></tr>
2438 <tr><td><tt>eq</tt></td><td>undefined</td><td>var1 == var2</td></tr>
2439 <tr><td><tt>ne</tt></td><td>undefined</td><td>var1 != var2</td></tr>
2440 <tr><td><tt>lt</tt></td><td>undefined</td><td>var1 &lt; var2</td></tr>
2441 <tr><td><tt>gt</tt></td><td>undefined</td><td>var1 &gt; var2</td></tr>
2442 <tr><td><tt>le</tt></td><td>undefined</td><td>var1 &lt;= var2</td></tr>
2443 <tr><td><tt>ge</tt></td><td>undefined</td><td>var1 &gt;= var2</td></tr>
2444 <tr><td><tt>oeq</tt></td><td>false</td><td>var1 == var2</td></tr>
2445 <tr><td><tt>one</tt></td><td>false</td><td>var1 != var2</td></tr>
2446 <tr><td><tt>olt</tt></td><td>false</td><td>var1 &lt; var2</td></tr>
2447 <tr><td><tt>ogt</tt></td><td>false</td><td>var1 &gt; var2</td></tr>
2448 <tr><td><tt>ole</tt></td><td>false</td><td>var1 &lt;= var2</td></tr>
2449 <tr><td><tt>oge</tt></td><td>false</td><td>var1 &gt;= var2</td></tr>
2450 <tr><td><tt>ueq</tt></td><td>true</td><td>var1 == var2</td></tr>
2451 <tr><td><tt>une</tt></td><td>true</td><td>var1 != var2</td></tr>
2452 <tr><td><tt>ult</tt></td><td>true</td><td>var1 &lt; var2</td></tr>
2453 <tr><td><tt>ugt</tt></td><td>true</td><td>var1 &gt; var2</td></tr>
2454 <tr><td><tt>ule</tt></td><td>true</td><td>var1 &lt;= var2</td></tr>
2455 <tr><td><tt>uge</tt></td><td>true</td><td>var1 &gt;= var2</td></tr>
2456 <tr><td><tt>o</tt></td><td>false</td><td>always</td></tr>
2457 <tr><td><tt>u</tt></td><td>true</td><td>never</td></tr>
2458 <tr><td><tt>true</tt></td><td>true</td><td>always</td></tr>
2459 <tr><td><tt>false</tt></td><td>false</td><td>never</td></tr>
2460 </tbody>
2461</table>
2462
2463<h5>Example:</h5>
Robert Bocchino230bd092006-03-14 23:22:57 +00002464<pre> &lt;result&gt; = vsetfp eq &lt;2 x float&gt; &lt;float 0.0, float 1.0&gt;, &lt;float 1.0, float 0.0&gt; <i>; yields {&lt;2 x bool&gt;}:result = false, false</i>
2465 &lt;result&gt; = vsetfp ne &lt;2 x float&gt; &lt;float 0.0, float 1.0&gt;, &lt;float 1.0, float 0.0&gt; <i>; yields {&lt;2 x bool&gt;}:result = true, true</i>
2466 &lt;result&gt; = vsetfp lt &lt;2 x float&gt; &lt;float 0.0, float 1.0&gt;, &lt;float 1.0, float 0.0&gt; <i>; yields {&lt;2 x bool&gt;}:result = true, false</i>
2467 &lt;result&gt; = vsetfp gt &lt;2 x float&gt; &lt;float 0.0, float 1.0&gt;, &lt;float 1.0, float 0.0&gt; <i>; yields {&lt;2 x bool&gt;}:result = false, true</i>
2468 &lt;result&gt; = vsetfp le &lt;2 x float&gt; &lt;float 0.0, float 1.0&gt;, &lt;float 1.0, float 0.0&gt; <i>; yields {&lt;2 x bool&gt;}:result = true, false</i>
2469 &lt;result&gt; = vsetfp ge &lt;2 x float&gt; &lt;float 0.0, float 1.0&gt;, &lt;float 1.0, float 0.0&gt; <i>; yields {&lt;2 x bool&gt;}:result = false, true</i>
Robert Bocchino7b81c752006-02-17 21:18:08 +00002470</pre>
2471</div>
2472
2473<!-- _______________________________________________________________________ -->
2474<div class="doc_subsubsection">
2475 <a name="i_vselect">'<tt>vselect</tt>' Instruction</a>
2476</div>
2477
2478<div class="doc_text">
2479
2480<h5>Syntax:</h5>
2481
2482<pre>
2483 &lt;result&gt; = vselect &lt;n x bool&gt; &lt;cond&gt;, &lt;n x &lt;ty&gt;&gt; &lt;val1&gt;, &lt;n x &lt;ty&gt;&gt; &lt;val2&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2484</pre>
2485
2486<h5>Overview:</h5>
2487
2488<p>
2489The '<tt>vselect</tt>' instruction chooses one value at each position
2490of a vector based on a condition.
2491</p>
2492
2493
2494<h5>Arguments:</h5>
2495
2496<p>
2497The '<tt>vselect</tt>' instruction requires a <a
2498href="#t_packed">packed</a> <tt>bool</tt> value indicating the
2499condition at each vector position, and two values of the same packed
2500type. All three operands must have the same length. The type of the
2501result is the same as the type of the two value operands.</p>
2502
2503<h5>Semantics:</h5>
2504
2505<p>
2506At each position where the <tt>bool</tt> vector is true, that position
2507of the result gets its value from the first value argument; otherwise,
2508it gets its value from the second value argument.
2509</p>
2510
2511<h5>Example:</h5>
2512
2513<pre>
2514 %X = vselect bool &lt;2 x bool&gt; &lt;bool true, bool false&gt;, &lt;2 x ubyte&gt; &lt;ubyte 17, ubyte 17&gt;,
2515 &lt;2 x ubyte&gt; &lt;ubyte 42, ubyte 42&gt; <i>; yields &lt;2 x ubyte&gt;:17, 42</i>
2516</pre>
2517</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002518
Robert Bocchino3a558662006-01-05 17:37:02 +00002519<!-- _______________________________________________________________________ -->
2520<div class="doc_subsubsection">
2521 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2522</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002523
Robert Bocchino3a558662006-01-05 17:37:02 +00002524<div class="doc_text">
2525
2526<h5>Syntax:</h5>
2527
2528<pre>
2529 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, uint &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2530</pre>
2531
2532<h5>Overview:</h5>
2533
2534<p>
2535The '<tt>extractelement</tt>' instruction extracts a single scalar
Robert Bocchino05ccd702006-01-15 20:48:27 +00002536element from a packed vector at a specified index.
Robert Bocchino3a558662006-01-05 17:37:02 +00002537</p>
2538
2539
2540<h5>Arguments:</h5>
2541
2542<p>
2543The first operand of an '<tt>extractelement</tt>' instruction is a
2544value of <a href="#t_packed">packed</a> type. The second operand is
2545an index indicating the position from which to extract the element.
2546The index may be a variable.</p>
2547
2548<h5>Semantics:</h5>
2549
2550<p>
2551The result is a scalar of the same type as the element type of
2552<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2553<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2554results are undefined.
2555</p>
2556
2557<h5>Example:</h5>
2558
2559<pre>
2560 %result = extractelement &lt;4 x int&gt; %vec, uint 0 <i>; yields int</i>
2561</pre>
2562</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002563
2564
Chris Lattner33ba0d92001-07-09 00:26:23 +00002565<!-- _______________________________________________________________________ -->
Chris Lattner2bff5242005-05-06 05:47:36 +00002566<div class="doc_subsubsection">
Robert Bocchino05ccd702006-01-15 20:48:27 +00002567 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2568</div>
2569
2570<div class="doc_text">
2571
2572<h5>Syntax:</h5>
2573
2574<pre>
2575 &lt;result&gt; = insertelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt, uint &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2576</pre>
2577
2578<h5>Overview:</h5>
2579
2580<p>
2581The '<tt>insertelement</tt>' instruction inserts a scalar
2582element into a packed vector at a specified index.
2583</p>
2584
2585
2586<h5>Arguments:</h5>
2587
2588<p>
2589The first operand of an '<tt>insertelement</tt>' instruction is a
2590value of <a href="#t_packed">packed</a> type. The second operand is a
2591scalar value whose type must equal the element type of the first
2592operand. The third operand is an index indicating the position at
2593which to insert the value. The index may be a variable.</p>
2594
2595<h5>Semantics:</h5>
2596
2597<p>
2598The result is a packed vector of the same type as <tt>val</tt>. Its
2599element values are those of <tt>val</tt> except at position
2600<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2601exceeds the length of <tt>val</tt>, the results are undefined.
2602</p>
2603
2604<h5>Example:</h5>
2605
2606<pre>
2607 %result = insertelement &lt;4 x int&gt; %vec, int 1, uint 0 <i>; yields &lt;4 x int&gt;</i>
2608</pre>
2609</div>
2610
2611
2612<!-- _______________________________________________________________________ -->
2613<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +00002614 <a name="i_call">'<tt>call</tt>' Instruction</a>
2615</div>
2616
Misha Brukman9d0919f2003-11-08 01:05:38 +00002617<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +00002618
Chris Lattner00950542001-06-06 20:29:01 +00002619<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002620<pre>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002621 &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 +00002622</pre>
2623
Chris Lattner00950542001-06-06 20:29:01 +00002624<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002625
Misha Brukman9d0919f2003-11-08 01:05:38 +00002626<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002627
Chris Lattner00950542001-06-06 20:29:01 +00002628<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002629
Misha Brukman9d0919f2003-11-08 01:05:38 +00002630<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002631
Chris Lattner6536cfe2002-05-06 22:08:29 +00002632<ol>
Chris Lattner261efe92003-11-25 01:02:51 +00002633 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002634 <p>The optional "tail" marker indicates whether the callee function accesses
2635 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +00002636 function call is eligible for tail call optimization. Note that calls may
2637 be marked "tail" even if they do not occur before a <a
2638 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +00002639 </li>
2640 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002641 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
2642 convention</a> the call should use. If none is specified, the call defaults
2643 to using C calling conventions.
2644 </li>
2645 <li>
Chris Lattner2bff5242005-05-06 05:47:36 +00002646 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
2647 being invoked. The argument types must match the types implied by this
John Criswellfc6b8952005-05-16 16:17:45 +00002648 signature. This type can be omitted if the function is not varargs and
2649 if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002650 </li>
2651 <li>
2652 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
2653 be invoked. In most cases, this is a direct function invocation, but
2654 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +00002655 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002656 </li>
2657 <li>
2658 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +00002659 function signature argument types. All arguments must be of
2660 <a href="#t_firstclass">first class</a> type. If the function signature
2661 indicates the function accepts a variable number of arguments, the extra
2662 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002663 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002664</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00002665
Chris Lattner00950542001-06-06 20:29:01 +00002666<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002667
Chris Lattner261efe92003-11-25 01:02:51 +00002668<p>The '<tt>call</tt>' instruction is used to cause control flow to
2669transfer to a specified function, with its incoming arguments bound to
2670the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
2671instruction in the called function, control flow continues with the
2672instruction after the function call, and the return value of the
2673function is bound to the result argument. This is a simpler case of
2674the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002675
Chris Lattner00950542001-06-06 20:29:01 +00002676<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002677
2678<pre>
2679 %retval = call int %test(int %argc)
2680 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
2681 %X = tail call int %foo()
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002682 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
Chris Lattner2bff5242005-05-06 05:47:36 +00002683</pre>
2684
Misha Brukman9d0919f2003-11-08 01:05:38 +00002685</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002686
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002687<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +00002688<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +00002689 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002690</div>
2691
Misha Brukman9d0919f2003-11-08 01:05:38 +00002692<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +00002693
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002694<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002695
2696<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002697 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00002698</pre>
2699
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002700<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002701
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002702<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +00002703the "variable argument" area of a function call. It is used to implement the
2704<tt>va_arg</tt> macro in C.</p>
2705
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002706<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002707
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002708<p>This instruction takes a <tt>va_list*</tt> value and the type of
2709the argument. It returns a value of the specified argument type and
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00002710increments the <tt>va_list</tt> to point to the next argument. Again, the
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002711actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002712
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002713<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002714
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002715<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
2716type from the specified <tt>va_list</tt> and causes the
2717<tt>va_list</tt> to point to the next argument. For more information,
2718see the variable argument handling <a href="#int_varargs">Intrinsic
2719Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002720
2721<p>It is legal for this instruction to be called in a function which does not
2722take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002723function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002724
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002725<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +00002726href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +00002727argument.</p>
2728
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002729<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002730
2731<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
2732
Misha Brukman9d0919f2003-11-08 01:05:38 +00002733</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002734
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002735<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00002736<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
2737<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002738
Misha Brukman9d0919f2003-11-08 01:05:38 +00002739<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +00002740
2741<p>LLVM supports the notion of an "intrinsic function". These functions have
John Criswellfc6b8952005-05-16 16:17:45 +00002742well known names and semantics and are required to follow certain
Chris Lattner33aec9e2004-02-12 17:01:32 +00002743restrictions. Overall, these instructions represent an extension mechanism for
2744the LLVM language that does not require changing all of the transformations in
2745LLVM to add to the language (or the bytecode reader/writer, the parser,
2746etc...).</p>
2747
John Criswellfc6b8952005-05-16 16:17:45 +00002748<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
2749prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
Chris Lattner33aec9e2004-02-12 17:01:32 +00002750this. Intrinsic functions must always be external functions: you cannot define
2751the body of intrinsic functions. Intrinsic functions may only be used in call
2752or invoke instructions: it is illegal to take the address of an intrinsic
2753function. Additionally, because intrinsic functions are part of the LLVM
2754language, it is required that they all be documented here if any are added.</p>
2755
2756
John Criswellfc6b8952005-05-16 16:17:45 +00002757<p>To learn how to add an intrinsic function, please see the <a
Chris Lattner590cff32005-05-11 03:35:57 +00002758href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +00002759</p>
2760
Misha Brukman9d0919f2003-11-08 01:05:38 +00002761</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002762
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002763<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002764<div class="doc_subsection">
2765 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
2766</div>
2767
Misha Brukman9d0919f2003-11-08 01:05:38 +00002768<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002769
Misha Brukman9d0919f2003-11-08 01:05:38 +00002770<p>Variable argument support is defined in LLVM with the <a
Chris Lattnerfb6977d2006-01-13 23:26:01 +00002771 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner261efe92003-11-25 01:02:51 +00002772intrinsic functions. These functions are related to the similarly
2773named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002774
Chris Lattner261efe92003-11-25 01:02:51 +00002775<p>All of these functions operate on arguments that use a
2776target-specific value type "<tt>va_list</tt>". The LLVM assembly
2777language reference manual does not define what this type is, so all
2778transformations should be prepared to handle intrinsics with any type
2779used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002780
Misha Brukman9d0919f2003-11-08 01:05:38 +00002781<p>This example shows how the <a href="#i_vanext"><tt>vanext</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +00002782instruction and the variable argument handling intrinsic functions are
2783used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002784
Chris Lattner33aec9e2004-02-12 17:01:32 +00002785<pre>
2786int %test(int %X, ...) {
2787 ; Initialize variable argument processing
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002788 %ap = alloca sbyte*
2789 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002790
2791 ; Read a single integer argument
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002792 %tmp = va_arg sbyte** %ap, int
Chris Lattner33aec9e2004-02-12 17:01:32 +00002793
2794 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002795 %aq = alloca sbyte*
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002796 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002797 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002798
2799 ; Stop processing of arguments.
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002800 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002801 ret int %tmp
2802}
2803</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002804</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002805
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002806<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002807<div class="doc_subsubsection">
2808 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
2809</div>
2810
2811
Misha Brukman9d0919f2003-11-08 01:05:38 +00002812<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002813<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002814<pre> declare void %llvm.va_start(&lt;va_list&gt;* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002815<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002816<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
2817<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
2818href="#i_va_arg">va_arg</a></tt>.</p>
2819
2820<h5>Arguments:</h5>
2821
2822<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
2823
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002824<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002825
2826<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
2827macro available in C. In a target-dependent way, it initializes the
2828<tt>va_list</tt> element the argument points to, so that the next call to
2829<tt>va_arg</tt> will produce the first variable argument passed to the function.
2830Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
2831last argument of the function, the compiler can figure that out.</p>
2832
Misha Brukman9d0919f2003-11-08 01:05:38 +00002833</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002834
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002835<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002836<div class="doc_subsubsection">
2837 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
2838</div>
2839
Misha Brukman9d0919f2003-11-08 01:05:38 +00002840<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002841<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002842<pre> declare void %llvm.va_end(&lt;va_list*&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002843<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002844<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
2845which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
2846or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002847<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002848<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002849<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002850<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002851macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
2852Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
2853 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
2854with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002855</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002856
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002857<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002858<div class="doc_subsubsection">
2859 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
2860</div>
2861
Misha Brukman9d0919f2003-11-08 01:05:38 +00002862<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002863
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002864<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002865
2866<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002867 declare void %llvm.va_copy(&lt;va_list&gt;* &lt;destarglist&gt;,
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002868 &lt;va_list&gt;* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00002869</pre>
2870
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002871<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002872
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002873<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
2874the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002875
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002876<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002877
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002878<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002879The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002880
Chris Lattnerd7923912004-05-23 21:06:01 +00002881
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002882<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002883
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002884<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
2885available in C. In a target-dependent way, it copies the source
2886<tt>va_list</tt> element into the destination list. This intrinsic is necessary
2887because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
Chris Lattnerd7923912004-05-23 21:06:01 +00002888arbitrarily complex and require memory allocation, for example.</p>
2889
Misha Brukman9d0919f2003-11-08 01:05:38 +00002890</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002891
Chris Lattner33aec9e2004-02-12 17:01:32 +00002892<!-- ======================================================================= -->
2893<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +00002894 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
2895</div>
2896
2897<div class="doc_text">
2898
2899<p>
2900LLVM support for <a href="GarbageCollection.html">Accurate Garbage
2901Collection</a> requires the implementation and generation of these intrinsics.
2902These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
2903stack</a>, as well as garbage collector implementations that require <a
2904href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
2905Front-ends for type-safe garbage collected languages should generate these
2906intrinsics to make use of the LLVM garbage collectors. For more details, see <a
2907href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
2908</p>
2909</div>
2910
2911<!-- _______________________________________________________________________ -->
2912<div class="doc_subsubsection">
2913 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
2914</div>
2915
2916<div class="doc_text">
2917
2918<h5>Syntax:</h5>
2919
2920<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002921 declare void %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00002922</pre>
2923
2924<h5>Overview:</h5>
2925
John Criswell9e2485c2004-12-10 15:51:16 +00002926<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +00002927the code generator, and allows some metadata to be associated with it.</p>
2928
2929<h5>Arguments:</h5>
2930
2931<p>The first argument specifies the address of a stack object that contains the
2932root pointer. The second pointer (which must be either a constant or a global
2933value address) contains the meta-data to be associated with the root.</p>
2934
2935<h5>Semantics:</h5>
2936
2937<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
2938location. At compile-time, the code generator generates information to allow
2939the runtime to find the pointer at GC safe points.
2940</p>
2941
2942</div>
2943
2944
2945<!-- _______________________________________________________________________ -->
2946<div class="doc_subsubsection">
2947 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
2948</div>
2949
2950<div class="doc_text">
2951
2952<h5>Syntax:</h5>
2953
2954<pre>
Chris Lattner80626e92006-03-14 20:02:51 +00002955 declare sbyte* %llvm.gcread(sbyte* %ObjPtr, sbyte** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00002956</pre>
2957
2958<h5>Overview:</h5>
2959
2960<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
2961locations, allowing garbage collector implementations that require read
2962barriers.</p>
2963
2964<h5>Arguments:</h5>
2965
Chris Lattner80626e92006-03-14 20:02:51 +00002966<p>The second argument is the address to read from, which should be an address
2967allocated from the garbage collector. The first object is a pointer to the
2968start of the referenced object, if needed by the language runtime (otherwise
2969null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002970
2971<h5>Semantics:</h5>
2972
2973<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
2974instruction, but may be replaced with substantially more complex code by the
2975garbage collector runtime, as needed.</p>
2976
2977</div>
2978
2979
2980<!-- _______________________________________________________________________ -->
2981<div class="doc_subsubsection">
2982 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
2983</div>
2984
2985<div class="doc_text">
2986
2987<h5>Syntax:</h5>
2988
2989<pre>
Chris Lattner80626e92006-03-14 20:02:51 +00002990 declare void %llvm.gcwrite(sbyte* %P1, sbyte* %Obj, sbyte** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00002991</pre>
2992
2993<h5>Overview:</h5>
2994
2995<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
2996locations, allowing garbage collector implementations that require write
2997barriers (such as generational or reference counting collectors).</p>
2998
2999<h5>Arguments:</h5>
3000
Chris Lattner80626e92006-03-14 20:02:51 +00003001<p>The first argument is the reference to store, the second is the start of the
3002object to store it to, and the third is the address of the field of Obj to
3003store to. If the runtime does not require a pointer to the object, Obj may be
3004null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00003005
3006<h5>Semantics:</h5>
3007
3008<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
3009instruction, but may be replaced with substantially more complex code by the
3010garbage collector runtime, as needed.</p>
3011
3012</div>
3013
3014
3015
3016<!-- ======================================================================= -->
3017<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +00003018 <a name="int_codegen">Code Generator Intrinsics</a>
3019</div>
3020
3021<div class="doc_text">
3022<p>
3023These intrinsics are provided by LLVM to expose special features that may only
3024be implemented with code generator support.
3025</p>
3026
3027</div>
3028
3029<!-- _______________________________________________________________________ -->
3030<div class="doc_subsubsection">
3031 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
3032</div>
3033
3034<div class="doc_text">
3035
3036<h5>Syntax:</h5>
3037<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00003038 declare sbyte *%llvm.returnaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003039</pre>
3040
3041<h5>Overview:</h5>
3042
3043<p>
3044The '<tt>llvm.returnaddress</tt>' intrinsic returns a target-specific value
3045indicating the return address of the current function or one of its callers.
3046</p>
3047
3048<h5>Arguments:</h5>
3049
3050<p>
3051The argument to this intrinsic indicates which function to return the address
3052for. Zero indicates the calling function, one indicates its caller, etc. The
3053argument is <b>required</b> to be a constant integer value.
3054</p>
3055
3056<h5>Semantics:</h5>
3057
3058<p>
3059The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
3060the return address of the specified call frame, or zero if it cannot be
3061identified. The value returned by this intrinsic is likely to be incorrect or 0
3062for arguments other than zero, so it should only be used for debugging purposes.
3063</p>
3064
3065<p>
3066Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00003067aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00003068source-language caller.
3069</p>
3070</div>
3071
3072
3073<!-- _______________________________________________________________________ -->
3074<div class="doc_subsubsection">
3075 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
3076</div>
3077
3078<div class="doc_text">
3079
3080<h5>Syntax:</h5>
3081<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00003082 declare sbyte *%llvm.frameaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003083</pre>
3084
3085<h5>Overview:</h5>
3086
3087<p>
3088The '<tt>llvm.frameaddress</tt>' intrinsic returns the target-specific frame
3089pointer value for the specified stack frame.
3090</p>
3091
3092<h5>Arguments:</h5>
3093
3094<p>
3095The argument to this intrinsic indicates which function to return the frame
3096pointer for. Zero indicates the calling function, one indicates its caller,
3097etc. The argument is <b>required</b> to be a constant integer value.
3098</p>
3099
3100<h5>Semantics:</h5>
3101
3102<p>
3103The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
3104the frame address of the specified call frame, or zero if it cannot be
3105identified. The value returned by this intrinsic is likely to be incorrect or 0
3106for arguments other than zero, so it should only be used for debugging purposes.
3107</p>
3108
3109<p>
3110Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00003111aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00003112source-language caller.
3113</p>
3114</div>
3115
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003116<!-- _______________________________________________________________________ -->
3117<div class="doc_subsubsection">
Chris Lattner57e1f392006-01-13 02:03:13 +00003118 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
3119</div>
3120
3121<div class="doc_text">
3122
3123<h5>Syntax:</h5>
3124<pre>
3125 declare sbyte *%llvm.stacksave()
3126</pre>
3127
3128<h5>Overview:</h5>
3129
3130<p>
3131The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
3132the function stack, for use with <a href="#i_stackrestore">
3133<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
3134features like scoped automatic variable sized arrays in C99.
3135</p>
3136
3137<h5>Semantics:</h5>
3138
3139<p>
3140This intrinsic returns a opaque pointer value that can be passed to <a
3141href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
3142<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
3143<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
3144state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
3145practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
3146that were allocated after the <tt>llvm.stacksave</tt> was executed.
3147</p>
3148
3149</div>
3150
3151<!-- _______________________________________________________________________ -->
3152<div class="doc_subsubsection">
3153 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
3154</div>
3155
3156<div class="doc_text">
3157
3158<h5>Syntax:</h5>
3159<pre>
3160 declare void %llvm.stackrestore(sbyte* %ptr)
3161</pre>
3162
3163<h5>Overview:</h5>
3164
3165<p>
3166The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
3167the function stack to the state it was in when the corresponding <a
3168href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
3169useful for implementing language features like scoped automatic variable sized
3170arrays in C99.
3171</p>
3172
3173<h5>Semantics:</h5>
3174
3175<p>
3176See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
3177</p>
3178
3179</div>
3180
3181
3182<!-- _______________________________________________________________________ -->
3183<div class="doc_subsubsection">
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003184 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
3185</div>
3186
3187<div class="doc_text">
3188
3189<h5>Syntax:</h5>
3190<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003191 declare void %llvm.prefetch(sbyte * &lt;address&gt;,
3192 uint &lt;rw&gt;, uint &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003193</pre>
3194
3195<h5>Overview:</h5>
3196
3197
3198<p>
3199The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +00003200a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
3201no
3202effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +00003203characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003204</p>
3205
3206<h5>Arguments:</h5>
3207
3208<p>
3209<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
3210determining if the fetch should be for a read (0) or write (1), and
3211<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +00003212locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003213<tt>locality</tt> arguments must be constant integers.
3214</p>
3215
3216<h5>Semantics:</h5>
3217
3218<p>
3219This intrinsic does not modify the behavior of the program. In particular,
3220prefetches cannot trap and do not produce a value. On targets that support this
3221intrinsic, the prefetch can provide hints to the processor cache for better
3222performance.
3223</p>
3224
3225</div>
3226
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003227<!-- _______________________________________________________________________ -->
3228<div class="doc_subsubsection">
3229 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
3230</div>
3231
3232<div class="doc_text">
3233
3234<h5>Syntax:</h5>
3235<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003236 declare void %llvm.pcmarker( uint &lt;id&gt; )
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003237</pre>
3238
3239<h5>Overview:</h5>
3240
3241
3242<p>
John Criswellfc6b8952005-05-16 16:17:45 +00003243The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
3244(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003245code to simulators and other tools. The method is target specific, but it is
3246expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00003247The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +00003248after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb3e7afd2006-03-24 07:16:10 +00003249optimizations. The intended use is to be inserted after optimizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +00003250correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003251</p>
3252
3253<h5>Arguments:</h5>
3254
3255<p>
3256<tt>id</tt> is a numerical id identifying the marker.
3257</p>
3258
3259<h5>Semantics:</h5>
3260
3261<p>
3262This intrinsic does not modify the behavior of the program. Backends that do not
3263support this intrinisic may ignore it.
3264</p>
3265
3266</div>
3267
Andrew Lenharth51b8d542005-11-11 16:47:30 +00003268<!-- _______________________________________________________________________ -->
3269<div class="doc_subsubsection">
3270 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
3271</div>
3272
3273<div class="doc_text">
3274
3275<h5>Syntax:</h5>
3276<pre>
3277 declare ulong %llvm.readcyclecounter( )
3278</pre>
3279
3280<h5>Overview:</h5>
3281
3282
3283<p>
3284The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
3285counter register (or similar low latency, high accuracy clocks) on those targets
3286that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
3287As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
3288should only be used for small timings.
3289</p>
3290
3291<h5>Semantics:</h5>
3292
3293<p>
3294When directly supported, reading the cycle counter should not modify any memory.
3295Implementations are allowed to either return a application specific value or a
3296system wide value. On backends without support, this is lowered to a constant 0.
3297</p>
3298
3299</div>
3300
Chris Lattner10610642004-02-14 04:08:35 +00003301<!-- ======================================================================= -->
3302<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +00003303 <a name="int_libc">Standard C Library Intrinsics</a>
3304</div>
3305
3306<div class="doc_text">
3307<p>
Chris Lattner10610642004-02-14 04:08:35 +00003308LLVM provides intrinsics for a few important standard C library functions.
3309These intrinsics allow source-language front-ends to pass information about the
3310alignment of the pointer arguments to the code generator, providing opportunity
3311for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003312</p>
3313
3314</div>
3315
3316<!-- _______________________________________________________________________ -->
3317<div class="doc_subsubsection">
3318 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
3319</div>
3320
3321<div class="doc_text">
3322
3323<h5>Syntax:</h5>
3324<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00003325 declare void %llvm.memcpy.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3326 uint &lt;len&gt;, uint &lt;align&gt;)
3327 declare void %llvm.memcpy.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3328 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003329</pre>
3330
3331<h5>Overview:</h5>
3332
3333<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003334The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00003335location to the destination location.
3336</p>
3337
3338<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003339Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
3340intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003341</p>
3342
3343<h5>Arguments:</h5>
3344
3345<p>
3346The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00003347the source. The third argument is an integer argument
Chris Lattner33aec9e2004-02-12 17:01:32 +00003348specifying the number of bytes to copy, and the fourth argument is the alignment
3349of the source and destination locations.
3350</p>
3351
Chris Lattner3301ced2004-02-12 21:18:15 +00003352<p>
3353If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00003354the caller guarantees that both the source and destination pointers are aligned
3355to that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00003356</p>
3357
Chris Lattner33aec9e2004-02-12 17:01:32 +00003358<h5>Semantics:</h5>
3359
3360<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003361The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00003362location to the destination location, which are not allowed to overlap. It
3363copies "len" bytes of memory over. If the argument is known to be aligned to
3364some boundary, this can be specified as the fourth argument, otherwise it should
3365be set to 0 or 1.
3366</p>
3367</div>
3368
3369
Chris Lattner0eb51b42004-02-12 18:10:10 +00003370<!-- _______________________________________________________________________ -->
3371<div class="doc_subsubsection">
3372 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
3373</div>
3374
3375<div class="doc_text">
3376
3377<h5>Syntax:</h5>
3378<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00003379 declare void %llvm.memmove.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3380 uint &lt;len&gt;, uint &lt;align&gt;)
3381 declare void %llvm.memmove.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3382 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00003383</pre>
3384
3385<h5>Overview:</h5>
3386
3387<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003388The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
3389location to the destination location. It is similar to the
3390'<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattner0eb51b42004-02-12 18:10:10 +00003391</p>
3392
3393<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003394Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
3395intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner0eb51b42004-02-12 18:10:10 +00003396</p>
3397
3398<h5>Arguments:</h5>
3399
3400<p>
3401The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00003402the source. The third argument is an integer argument
Chris Lattner0eb51b42004-02-12 18:10:10 +00003403specifying the number of bytes to copy, and the fourth argument is the alignment
3404of the source and destination locations.
3405</p>
3406
Chris Lattner3301ced2004-02-12 21:18:15 +00003407<p>
3408If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00003409the caller guarantees that the source and destination pointers are aligned to
3410that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00003411</p>
3412
Chris Lattner0eb51b42004-02-12 18:10:10 +00003413<h5>Semantics:</h5>
3414
3415<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003416The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner0eb51b42004-02-12 18:10:10 +00003417location to the destination location, which may overlap. It
3418copies "len" bytes of memory over. If the argument is known to be aligned to
3419some boundary, this can be specified as the fourth argument, otherwise it should
3420be set to 0 or 1.
3421</p>
3422</div>
3423
Chris Lattner8ff75902004-01-06 05:31:32 +00003424
Chris Lattner10610642004-02-14 04:08:35 +00003425<!-- _______________________________________________________________________ -->
3426<div class="doc_subsubsection">
Chris Lattner5b310c32006-03-03 00:07:20 +00003427 <a name="i_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner10610642004-02-14 04:08:35 +00003428</div>
3429
3430<div class="doc_text">
3431
3432<h5>Syntax:</h5>
3433<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00003434 declare void %llvm.memset.i32(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
3435 uint &lt;len&gt;, uint &lt;align&gt;)
3436 declare void %llvm.memset.i64(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
3437 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003438</pre>
3439
3440<h5>Overview:</h5>
3441
3442<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003443The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner10610642004-02-14 04:08:35 +00003444byte value.
3445</p>
3446
3447<p>
3448Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
3449does not return a value, and takes an extra alignment argument.
3450</p>
3451
3452<h5>Arguments:</h5>
3453
3454<p>
3455The first argument is a pointer to the destination to fill, the second is the
Chris Lattner5b310c32006-03-03 00:07:20 +00003456byte value to fill it with, the third argument is an integer
Chris Lattner10610642004-02-14 04:08:35 +00003457argument specifying the number of bytes to fill, and the fourth argument is the
3458known alignment of destination location.
3459</p>
3460
3461<p>
3462If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00003463the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner10610642004-02-14 04:08:35 +00003464</p>
3465
3466<h5>Semantics:</h5>
3467
3468<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003469The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
3470the
Chris Lattner10610642004-02-14 04:08:35 +00003471destination location. If the argument is known to be aligned to some boundary,
3472this can be specified as the fourth argument, otherwise it should be set to 0 or
34731.
3474</p>
3475</div>
3476
3477
Chris Lattner32006282004-06-11 02:28:03 +00003478<!-- _______________________________________________________________________ -->
3479<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +00003480 <a name="i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a>
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003481</div>
3482
3483<div class="doc_text">
3484
3485<h5>Syntax:</h5>
3486<pre>
Reid Spencer0b118202006-01-16 21:12:35 +00003487 declare bool %llvm.isunordered.f32(float Val1, float Val2)
3488 declare bool %llvm.isunordered.f64(double Val1, double Val2)
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003489</pre>
3490
3491<h5>Overview:</h5>
3492
3493<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003494The '<tt>llvm.isunordered</tt>' intrinsics return true if either or both of the
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003495specified floating point values is a NAN.
3496</p>
3497
3498<h5>Arguments:</h5>
3499
3500<p>
3501The arguments are floating point numbers of the same type.
3502</p>
3503
3504<h5>Semantics:</h5>
3505
3506<p>
3507If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
3508false.
3509</p>
3510</div>
3511
3512
Chris Lattnera4d74142005-07-21 01:29:16 +00003513<!-- _______________________________________________________________________ -->
3514<div class="doc_subsubsection">
Chris Lattnerec6cb612006-01-16 22:38:59 +00003515 <a name="i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +00003516</div>
3517
3518<div class="doc_text">
3519
3520<h5>Syntax:</h5>
3521<pre>
Reid Spencer0b118202006-01-16 21:12:35 +00003522 declare double %llvm.sqrt.f32(float Val)
3523 declare double %llvm.sqrt.f64(double Val)
Chris Lattnera4d74142005-07-21 01:29:16 +00003524</pre>
3525
3526<h5>Overview:</h5>
3527
3528<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003529The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Chris Lattnera4d74142005-07-21 01:29:16 +00003530returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
3531<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
3532negative numbers (which allows for better optimization).
3533</p>
3534
3535<h5>Arguments:</h5>
3536
3537<p>
3538The argument and return value are floating point numbers of the same type.
3539</p>
3540
3541<h5>Semantics:</h5>
3542
3543<p>
3544This function returns the sqrt of the specified operand if it is a positive
3545floating point number.
3546</p>
3547</div>
3548
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003549<!-- ======================================================================= -->
3550<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +00003551 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003552</div>
3553
3554<div class="doc_text">
3555<p>
Nate Begeman7e36c472006-01-13 23:26:38 +00003556LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003557These allow efficient code generation for some algorithms.
3558</p>
3559
3560</div>
3561
3562<!-- _______________________________________________________________________ -->
3563<div class="doc_subsubsection">
Nate Begeman7e36c472006-01-13 23:26:38 +00003564 <a name="i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
3565</div>
3566
3567<div class="doc_text">
3568
3569<h5>Syntax:</h5>
3570<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003571 declare ushort %llvm.bswap.i16(ushort &lt;id&gt;)
3572 declare uint %llvm.bswap.i32(uint &lt;id&gt;)
3573 declare ulong %llvm.bswap.i64(ulong &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +00003574</pre>
3575
3576<h5>Overview:</h5>
3577
3578<p>
3579The '<tt>llvm.bwsap</tt>' family of intrinsics is used to byteswap a 16, 32 or
358064 bit quantity. These are useful for performing operations on data that is not
3581in the target's native byte order.
3582</p>
3583
3584<h5>Semantics:</h5>
3585
3586<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003587The <tt>llvm.bswap.16</tt> intrinsic returns a ushort value that has the high and low
3588byte of the input ushort swapped. Similarly, the <tt>llvm.bswap.i32</tt> intrinsic
Nate Begeman7e36c472006-01-13 23:26:38 +00003589returns a uint value that has the four bytes of the input uint swapped, so that
3590if the input bytes are numbered 0, 1, 2, 3 then the returned uint will have its
Chris Lattnerec6cb612006-01-16 22:38:59 +00003591bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i64</tt> intrinsic extends this concept
Nate Begeman7e36c472006-01-13 23:26:38 +00003592to 64 bits.
3593</p>
3594
3595</div>
3596
3597<!-- _______________________________________________________________________ -->
3598<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +00003599 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003600</div>
3601
3602<div class="doc_text">
3603
3604<h5>Syntax:</h5>
3605<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003606 declare ubyte %llvm.ctpop.i8 (ubyte &lt;src&gt;)
3607 declare ushort %llvm.ctpop.i16(ushort &lt;src&gt;)
3608 declare uint %llvm.ctpop.i32(uint &lt;src&gt;)
3609 declare ulong %llvm.ctpop.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003610</pre>
3611
3612<h5>Overview:</h5>
3613
3614<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003615The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
3616value.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003617</p>
3618
3619<h5>Arguments:</h5>
3620
3621<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003622The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00003623unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003624</p>
3625
3626<h5>Semantics:</h5>
3627
3628<p>
3629The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
3630</p>
3631</div>
3632
3633<!-- _______________________________________________________________________ -->
3634<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00003635 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003636</div>
3637
3638<div class="doc_text">
3639
3640<h5>Syntax:</h5>
3641<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003642 declare ubyte %llvm.ctlz.i8 (ubyte &lt;src&gt;)
3643 declare ushort %llvm.ctlz.i16(ushort &lt;src&gt;)
3644 declare uint %llvm.ctlz.i32(uint &lt;src&gt;)
3645 declare ulong %llvm.ctlz.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003646</pre>
3647
3648<h5>Overview:</h5>
3649
3650<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003651The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
3652leading zeros in a variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003653</p>
3654
3655<h5>Arguments:</h5>
3656
3657<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003658The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00003659unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003660</p>
3661
3662<h5>Semantics:</h5>
3663
3664<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00003665The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
3666in a variable. If the src == 0 then the result is the size in bits of the type
3667of src. For example, <tt>llvm.cttz(int 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003668</p>
3669</div>
Chris Lattner32006282004-06-11 02:28:03 +00003670
3671
Chris Lattnereff29ab2005-05-15 19:39:26 +00003672
3673<!-- _______________________________________________________________________ -->
3674<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00003675 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +00003676</div>
3677
3678<div class="doc_text">
3679
3680<h5>Syntax:</h5>
3681<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003682 declare ubyte %llvm.cttz.i8 (ubyte &lt;src&gt;)
3683 declare ushort %llvm.cttz.i16(ushort &lt;src&gt;)
3684 declare uint %llvm.cttz.i32(uint &lt;src&gt;)
3685 declare ulong %llvm.cttz.i64(ulong &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +00003686</pre>
3687
3688<h5>Overview:</h5>
3689
3690<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003691The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
3692trailing zeros.
Chris Lattnereff29ab2005-05-15 19:39:26 +00003693</p>
3694
3695<h5>Arguments:</h5>
3696
3697<p>
3698The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00003699unsigned integer type. The return type must match the argument type.
Chris Lattnereff29ab2005-05-15 19:39:26 +00003700</p>
3701
3702<h5>Semantics:</h5>
3703
3704<p>
3705The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
3706in a variable. If the src == 0 then the result is the size in bits of the type
3707of src. For example, <tt>llvm.cttz(2) = 1</tt>.
3708</p>
3709</div>
3710
Chris Lattner8ff75902004-01-06 05:31:32 +00003711<!-- ======================================================================= -->
3712<div class="doc_subsection">
3713 <a name="int_debugger">Debugger Intrinsics</a>
3714</div>
3715
3716<div class="doc_text">
3717<p>
3718The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
3719are described in the <a
3720href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
3721Debugging</a> document.
3722</p>
3723</div>
3724
3725
Chris Lattner00950542001-06-06 20:29:01 +00003726<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00003727<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00003728<address>
3729 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3730 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3731 <a href="http://validator.w3.org/check/referer"><img
3732 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3733
3734 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +00003735 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00003736 Last modified: $Date$
3737</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003738</body>
3739</html>