blob: 2be38d42a89cd8da9be1620590494235e699c623 [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 Lattner3df241e2006-04-08 23:07:04 +000094 <li><a href="#vectorops">Vector Operations</a>
95 <ol>
96 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
97 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
98 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +000099 </ol>
100 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000101 <li><a href="#memoryops">Memory Access Operations</a>
102 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000103 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
104 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
105 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000106 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
107 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
108 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000109 </ol>
110 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000111 <li><a href="#otherops">Other Operations</a>
112 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000113 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000114 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000115 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000116 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000117 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000118 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000119 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000120 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000121 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000122 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000123 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000124 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
125 <ol>
126 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
127 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
128 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
129 </ol>
130 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000131 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
132 <ol>
133 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
134 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
135 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
136 </ol>
137 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000138 <li><a href="#int_codegen">Code Generator Intrinsics</a>
139 <ol>
140 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
141 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
Chris Lattner57e1f392006-01-13 02:03:13 +0000142 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
143 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000144 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000145 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000146 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000147 </ol>
148 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000149 <li><a href="#int_libc">Standard C Library Intrinsics</a>
150 <ol>
Chris Lattner5b310c32006-03-03 00:07:20 +0000151 <li><a href="#i_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
152 <li><a href="#i_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
153 <li><a href="#i_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000154 <li><a href="#i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a></li>
155 <li><a href="#i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
Chris Lattnera4d74142005-07-21 01:29:16 +0000156
Chris Lattner33aec9e2004-02-12 17:01:32 +0000157 </ol>
158 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000159 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000160 <ol>
Nate Begeman7e36c472006-01-13 23:26:38 +0000161 <li><a href="#i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000162 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
163 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
164 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000165 </ol>
166 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000167 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000168 </ol>
169 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000170</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000171
172<div class="doc_author">
173 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
174 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000175</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000176
Chris Lattner00950542001-06-06 20:29:01 +0000177<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000178<div class="doc_section"> <a name="abstract">Abstract </a></div>
179<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000180
Misha Brukman9d0919f2003-11-08 01:05:38 +0000181<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000182<p>This document is a reference manual for the LLVM assembly language.
183LLVM is an SSA based representation that provides type safety,
184low-level operations, flexibility, and the capability of representing
185'all' high-level languages cleanly. It is the common code
186representation used throughout all phases of the LLVM compilation
187strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000188</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000189
Chris Lattner00950542001-06-06 20:29:01 +0000190<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000191<div class="doc_section"> <a name="introduction">Introduction</a> </div>
192<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000193
Misha Brukman9d0919f2003-11-08 01:05:38 +0000194<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000195
Chris Lattner261efe92003-11-25 01:02:51 +0000196<p>The LLVM code representation is designed to be used in three
197different forms: as an in-memory compiler IR, as an on-disk bytecode
198representation (suitable for fast loading by a Just-In-Time compiler),
199and as a human readable assembly language representation. This allows
200LLVM to provide a powerful intermediate representation for efficient
201compiler transformations and analysis, while providing a natural means
202to debug and visualize the transformations. The three different forms
203of LLVM are all equivalent. This document describes the human readable
204representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000205
John Criswellc1f786c2005-05-13 22:25:59 +0000206<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000207while being expressive, typed, and extensible at the same time. It
208aims to be a "universal IR" of sorts, by being at a low enough level
209that high-level ideas may be cleanly mapped to it (similar to how
210microprocessors are "universal IR's", allowing many source languages to
211be mapped to them). By providing type information, LLVM can be used as
212the target of optimizations: for example, through pointer analysis, it
213can be proven that a C automatic variable is never accessed outside of
214the current function... allowing it to be promoted to a simple SSA
215value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000216
Misha Brukman9d0919f2003-11-08 01:05:38 +0000217</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000218
Chris Lattner00950542001-06-06 20:29:01 +0000219<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000220<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000221
Misha Brukman9d0919f2003-11-08 01:05:38 +0000222<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000223
Chris Lattner261efe92003-11-25 01:02:51 +0000224<p>It is important to note that this document describes 'well formed'
225LLVM assembly language. There is a difference between what the parser
226accepts and what is considered 'well formed'. For example, the
227following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000228
229<pre>
230 %x = <a href="#i_add">add</a> int 1, %x
231</pre>
232
Chris Lattner261efe92003-11-25 01:02:51 +0000233<p>...because the definition of <tt>%x</tt> does not dominate all of
234its uses. The LLVM infrastructure provides a verification pass that may
235be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000236automatically run by the parser after parsing input assembly and by
Chris Lattner261efe92003-11-25 01:02:51 +0000237the optimizer before it outputs bytecode. The violations pointed out
238by the verifier pass indicate bugs in transformation passes or input to
239the parser.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000240
Chris Lattner261efe92003-11-25 01:02:51 +0000241<!-- Describe the typesetting conventions here. --> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000242
Chris Lattner00950542001-06-06 20:29:01 +0000243<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000244<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000245<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000246
Misha Brukman9d0919f2003-11-08 01:05:38 +0000247<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000248
Chris Lattner261efe92003-11-25 01:02:51 +0000249<p>LLVM uses three different forms of identifiers, for different
250purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000251
Chris Lattner00950542001-06-06 20:29:01 +0000252<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000253 <li>Named values are represented as a string of characters with a '%' prefix.
254 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
255 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
256 Identifiers which require other characters in their names can be surrounded
257 with quotes. In this way, anything except a <tt>"</tt> character can be used
258 in a name.</li>
259
260 <li>Unnamed values are represented as an unsigned numeric value with a '%'
261 prefix. For example, %12, %2, %44.</li>
262
Reid Spencercc16dc32004-12-09 18:02:53 +0000263 <li>Constants, which are described in a <a href="#constants">section about
264 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000265</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000266
267<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
268don't need to worry about name clashes with reserved words, and the set of
269reserved words may be expanded in the future without penalty. Additionally,
270unnamed identifiers allow a compiler to quickly come up with a temporary
271variable without having to avoid symbol table conflicts.</p>
272
Chris Lattner261efe92003-11-25 01:02:51 +0000273<p>Reserved words in LLVM are very similar to reserved words in other
274languages. There are keywords for different opcodes ('<tt><a
Chris Lattnere5d947b2004-12-09 16:36:40 +0000275href="#i_add">add</a></tt>', '<tt><a href="#i_cast">cast</a></tt>', '<tt><a
276href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
277href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
278and others. These reserved words cannot conflict with variable names, because
279none of them start with a '%' character.</p>
280
281<p>Here is an example of LLVM code to multiply the integer variable
282'<tt>%X</tt>' by 8:</p>
283
Misha Brukman9d0919f2003-11-08 01:05:38 +0000284<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000285
286<pre>
287 %result = <a href="#i_mul">mul</a> uint %X, 8
288</pre>
289
Misha Brukman9d0919f2003-11-08 01:05:38 +0000290<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000291
292<pre>
293 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
294</pre>
295
Misha Brukman9d0919f2003-11-08 01:05:38 +0000296<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000297
298<pre>
299 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
300 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
301 %result = <a href="#i_add">add</a> uint %1, %1
302</pre>
303
Chris Lattner261efe92003-11-25 01:02:51 +0000304<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
305important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000306
Chris Lattner00950542001-06-06 20:29:01 +0000307<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000308
309 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
310 line.</li>
311
312 <li>Unnamed temporaries are created when the result of a computation is not
313 assigned to a named value.</li>
314
Misha Brukman9d0919f2003-11-08 01:05:38 +0000315 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000316
Misha Brukman9d0919f2003-11-08 01:05:38 +0000317</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000318
John Criswelle4c57cc2005-05-12 16:52:32 +0000319<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000320demonstrating instructions, we will follow an instruction with a comment that
321defines the type and name of value produced. Comments are shown in italic
322text.</p>
323
Misha Brukman9d0919f2003-11-08 01:05:38 +0000324</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000325
326<!-- *********************************************************************** -->
327<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
328<!-- *********************************************************************** -->
329
330<!-- ======================================================================= -->
331<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
332</div>
333
334<div class="doc_text">
335
336<p>LLVM programs are composed of "Module"s, each of which is a
337translation unit of the input programs. Each module consists of
338functions, global variables, and symbol table entries. Modules may be
339combined together with the LLVM linker, which merges function (and
340global variable) definitions, resolves forward declarations, and merges
341symbol table entries. Here is an example of the "hello world" module:</p>
342
343<pre><i>; Declare the string constant as a global constant...</i>
344<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
345 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
346
347<i>; External declaration of the puts function</i>
348<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
349
350<i>; Definition of main function</i>
351int %main() { <i>; int()* </i>
352 <i>; Convert [13x sbyte]* to sbyte *...</i>
353 %cast210 = <a
354 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
355
356 <i>; Call puts function to write out the string to stdout...</i>
357 <a
358 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
359 <a
360 href="#i_ret">ret</a> int 0<br>}<br></pre>
361
362<p>This example is made up of a <a href="#globalvars">global variable</a>
363named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
364function, and a <a href="#functionstructure">function definition</a>
365for "<tt>main</tt>".</p>
366
Chris Lattnere5d947b2004-12-09 16:36:40 +0000367<p>In general, a module is made up of a list of global values,
368where both functions and global variables are global values. Global values are
369represented by a pointer to a memory location (in this case, a pointer to an
370array of char, and a pointer to a function), and have one of the following <a
371href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000372
Chris Lattnere5d947b2004-12-09 16:36:40 +0000373</div>
374
375<!-- ======================================================================= -->
376<div class="doc_subsection">
377 <a name="linkage">Linkage Types</a>
378</div>
379
380<div class="doc_text">
381
382<p>
383All Global Variables and Functions have one of the following types of linkage:
384</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000385
386<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000387
Chris Lattnerfa730212004-12-09 16:11:40 +0000388 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000389
390 <dd>Global values with internal linkage are only directly accessible by
391 objects in the current module. In particular, linking code into a module with
392 an internal global value may cause the internal to be renamed as necessary to
393 avoid collisions. Because the symbol is internal to the module, all
394 references can be updated. This corresponds to the notion of the
395 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattnerfa730212004-12-09 16:11:40 +0000396 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000397
Chris Lattnerfa730212004-12-09 16:11:40 +0000398 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000399
400 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
401 the twist that linking together two modules defining the same
402 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
403 is typically used to implement inline functions. Unreferenced
404 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000405 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000406
Chris Lattnerfa730212004-12-09 16:11:40 +0000407 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000408
409 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
410 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
411 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000412 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000413
Chris Lattnerfa730212004-12-09 16:11:40 +0000414 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000415
416 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
417 pointer to array type. When two global variables with appending linkage are
418 linked together, the two global arrays are appended together. This is the
419 LLVM, typesafe, equivalent of having the system linker append together
420 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000421 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000422
Chris Lattnerfa730212004-12-09 16:11:40 +0000423 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000424
425 <dd>If none of the above identifiers are used, the global is externally
426 visible, meaning that it participates in linkage and can be used to resolve
427 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000428 </dd>
429</dl>
430
Chris Lattnerfa730212004-12-09 16:11:40 +0000431<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
432variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
433variable and was linked with this one, one of the two would be renamed,
434preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
435external (i.e., lacking any linkage declarations), they are accessible
436outside of the current module. It is illegal for a function <i>declaration</i>
437to have any linkage type other than "externally visible".</a></p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000438
Chris Lattnerfa730212004-12-09 16:11:40 +0000439</div>
440
441<!-- ======================================================================= -->
442<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000443 <a name="callingconv">Calling Conventions</a>
444</div>
445
446<div class="doc_text">
447
448<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
449and <a href="#i_invoke">invokes</a> can all have an optional calling convention
450specified for the call. The calling convention of any pair of dynamic
451caller/callee must match, or the behavior of the program is undefined. The
452following calling conventions are supported by LLVM, and more may be added in
453the future:</p>
454
455<dl>
456 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
457
458 <dd>This calling convention (the default if no other calling convention is
459 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000460 supports varargs function calls and tolerates some mismatch in the declared
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000461 prototype and implemented declaration of the function (as does normal C).
462 </dd>
463
464 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
465
466 <dd>This calling convention attempts to make calls as fast as possible
467 (e.g. by passing things in registers). This calling convention allows the
468 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000469 without having to conform to an externally specified ABI. Implementations of
470 this convention should allow arbitrary tail call optimization to be supported.
471 This calling convention does not support varargs and requires the prototype of
472 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000473 </dd>
474
475 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
476
477 <dd>This calling convention attempts to make code in the caller as efficient
478 as possible under the assumption that the call is not commonly executed. As
479 such, these calls often preserve all registers so that the call does not break
480 any live ranges in the caller side. This calling convention does not support
481 varargs and requires the prototype of all callees to exactly match the
482 prototype of the function definition.
483 </dd>
484
Chris Lattnercfe6b372005-05-07 01:46:40 +0000485 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000486
487 <dd>Any calling convention may be specified by number, allowing
488 target-specific calling conventions to be used. Target specific calling
489 conventions start at 64.
490 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000491</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000492
493<p>More calling conventions can be added/defined on an as-needed basis, to
494support pascal conventions or any other well-known target-independent
495convention.</p>
496
497</div>
498
499<!-- ======================================================================= -->
500<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000501 <a name="globalvars">Global Variables</a>
502</div>
503
504<div class="doc_text">
505
Chris Lattner3689a342005-02-12 19:30:21 +0000506<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000507instead of run-time. Global variables may optionally be initialized, may have
508an explicit section to be placed in, and may
Chris Lattner2cbdc452005-11-06 08:02:57 +0000509have an optional explicit alignment specified. A
John Criswell0ec250c2005-10-24 16:17:18 +0000510variable may be defined as a global "constant," which indicates that the
Chris Lattner3689a342005-02-12 19:30:21 +0000511contents of the variable will <b>never</b> be modified (enabling better
512optimization, allowing the global data to be placed in the read-only section of
513an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000514cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000515
516<p>
517LLVM explicitly allows <em>declarations</em> of global variables to be marked
518constant, even if the final definition of the global is not. This capability
519can be used to enable slightly better optimization of the program, but requires
520the language definition to guarantee that optimizations based on the
521'constantness' are valid for the translation units that do not include the
522definition.
523</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000524
525<p>As SSA values, global variables define pointer values that are in
526scope (i.e. they dominate) all basic blocks in the program. Global
527variables always define a pointer to their "content" type because they
528describe a region of memory, and all memory objects in LLVM are
529accessed through pointers.</p>
530
Chris Lattner88f6c462005-11-12 00:45:07 +0000531<p>LLVM allows an explicit section to be specified for globals. If the target
532supports it, it will emit globals to the section specified.</p>
533
Chris Lattner2cbdc452005-11-06 08:02:57 +0000534<p>An explicit alignment may be specified for a global. If not present, or if
535the alignment is set to zero, the alignment of the global is set by the target
536to whatever it feels convenient. If an explicit alignment is specified, the
537global is forced to have at least that much alignment. All alignments must be
538a power of 2.</p>
539
Chris Lattnerfa730212004-12-09 16:11:40 +0000540</div>
541
542
543<!-- ======================================================================= -->
544<div class="doc_subsection">
545 <a name="functionstructure">Functions</a>
546</div>
547
548<div class="doc_text">
549
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000550<p>LLVM function definitions consist of an optional <a href="#linkage">linkage
551type</a>, an optional <a href="#callingconv">calling convention</a>, a return
Chris Lattner88f6c462005-11-12 00:45:07 +0000552type, a function name, a (possibly empty) argument list, an optional section,
553an optional alignment, an opening curly brace,
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000554a list of basic blocks, and a closing curly brace. LLVM function declarations
555are defined with the "<tt>declare</tt>" keyword, an optional <a
Chris Lattner2cbdc452005-11-06 08:02:57 +0000556href="#callingconv">calling convention</a>, a return type, a function name,
557a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000558
559<p>A function definition contains a list of basic blocks, forming the CFG for
560the function. Each basic block may optionally start with a label (giving the
561basic block a symbol table entry), contains a list of instructions, and ends
562with a <a href="#terminators">terminator</a> instruction (such as a branch or
563function return).</p>
564
John Criswelle4c57cc2005-05-12 16:52:32 +0000565<p>The first basic block in a program is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000566executed on entrance to the function, and it is not allowed to have predecessor
567basic blocks (i.e. there can not be any branches to the entry block of a
568function). Because the block can have no predecessors, it also cannot have any
569<a href="#i_phi">PHI nodes</a>.</p>
570
571<p>LLVM functions are identified by their name and type signature. Hence, two
572functions with the same name but different parameter lists or return values are
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000573considered different functions, and LLVM will resolve references to each
Chris Lattnerfa730212004-12-09 16:11:40 +0000574appropriately.</p>
575
Chris Lattner88f6c462005-11-12 00:45:07 +0000576<p>LLVM allows an explicit section to be specified for functions. If the target
577supports it, it will emit functions to the section specified.</p>
578
Chris Lattner2cbdc452005-11-06 08:02:57 +0000579<p>An explicit alignment may be specified for a function. If not present, or if
580the alignment is set to zero, the alignment of the function is set by the target
581to whatever it feels convenient. If an explicit alignment is specified, the
582function is forced to have at least that much alignment. All alignments must be
583a power of 2.</p>
584
Chris Lattnerfa730212004-12-09 16:11:40 +0000585</div>
586
Chris Lattner4e9aba72006-01-23 23:23:47 +0000587<!-- ======================================================================= -->
588<div class="doc_subsection">
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000589 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000590</div>
591
592<div class="doc_text">
593<p>
594Modules may contain "module-level inline asm" blocks, which corresponds to the
595GCC "file scope inline asm" blocks. These blocks are internally concatenated by
596LLVM and treated as a single unit, but may be separated in the .ll file if
597desired. The syntax is very simple:
598</p>
599
600<div class="doc_code"><pre>
Chris Lattner52599e12006-01-24 00:37:20 +0000601 module asm "inline asm code goes here"
602 module asm "more can go here"
Chris Lattner4e9aba72006-01-23 23:23:47 +0000603</pre></div>
604
605<p>The strings can contain any character by escaping non-printable characters.
606 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
607 for the number.
608</p>
609
610<p>
611 The inline asm code is simply printed to the machine code .s file when
612 assembly code is generated.
613</p>
614</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000615
616
Chris Lattner00950542001-06-06 20:29:01 +0000617<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000618<div class="doc_section"> <a name="typesystem">Type System</a> </div>
619<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000620
Misha Brukman9d0919f2003-11-08 01:05:38 +0000621<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000622
Misha Brukman9d0919f2003-11-08 01:05:38 +0000623<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000624intermediate representation. Being typed enables a number of
625optimizations to be performed on the IR directly, without having to do
626extra analyses on the side before the transformation. A strong type
627system makes it easier to read the generated code and enables novel
628analyses and transformations that are not feasible to perform on normal
629three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000630
631</div>
632
Chris Lattner00950542001-06-06 20:29:01 +0000633<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000634<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000635<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000636<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000637system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000638
Reid Spencerd3f876c2004-11-01 08:19:36 +0000639<table class="layout">
640 <tr class="layout">
641 <td class="left">
642 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000643 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000644 <tr><th>Type</th><th>Description</th></tr>
645 <tr><td><tt>void</tt></td><td>No value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000646 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
647 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
648 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
649 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
650 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000651 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000652 </tbody>
653 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000654 </td>
655 <td class="right">
656 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000657 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000658 <tr><th>Type</th><th>Description</th></tr>
659 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000660 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
661 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
662 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
663 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
664 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000665 </tbody>
666 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000667 </td>
668 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000669</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000670</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000671
Chris Lattner00950542001-06-06 20:29:01 +0000672<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000673<div class="doc_subsubsection"> <a name="t_classifications">Type
674Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000675<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000676<p>These different primitive types fall into a few useful
677classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000678
679<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000680 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000681 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000682 <tr>
683 <td><a name="t_signed">signed</a></td>
684 <td><tt>sbyte, short, int, long, float, double</tt></td>
685 </tr>
686 <tr>
687 <td><a name="t_unsigned">unsigned</a></td>
688 <td><tt>ubyte, ushort, uint, ulong</tt></td>
689 </tr>
690 <tr>
691 <td><a name="t_integer">integer</a></td>
692 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
693 </tr>
694 <tr>
695 <td><a name="t_integral">integral</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000696 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
697 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000698 </tr>
699 <tr>
700 <td><a name="t_floating">floating point</a></td>
701 <td><tt>float, double</tt></td>
702 </tr>
703 <tr>
704 <td><a name="t_firstclass">first class</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000705 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
706 float, double, <a href="#t_pointer">pointer</a>,
707 <a href="#t_packed">packed</a></tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000708 </tr>
709 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000710</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000711
Chris Lattner261efe92003-11-25 01:02:51 +0000712<p>The <a href="#t_firstclass">first class</a> types are perhaps the
713most important. Values of these types are the only ones which can be
714produced by instructions, passed as arguments, or used as operands to
715instructions. This means that all structures and arrays must be
716manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000717</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000718
Chris Lattner00950542001-06-06 20:29:01 +0000719<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000720<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000721
Misha Brukman9d0919f2003-11-08 01:05:38 +0000722<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000723
Chris Lattner261efe92003-11-25 01:02:51 +0000724<p>The real power in LLVM comes from the derived types in the system.
725This is what allows a programmer to represent arrays, functions,
726pointers, and other useful types. Note that these derived types may be
727recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000728
Misha Brukman9d0919f2003-11-08 01:05:38 +0000729</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000730
Chris Lattner00950542001-06-06 20:29:01 +0000731<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000732<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000733
Misha Brukman9d0919f2003-11-08 01:05:38 +0000734<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000735
Chris Lattner00950542001-06-06 20:29:01 +0000736<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000737
Misha Brukman9d0919f2003-11-08 01:05:38 +0000738<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000739sequentially in memory. The array type requires a size (number of
740elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000741
Chris Lattner7faa8832002-04-14 06:13:44 +0000742<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000743
744<pre>
745 [&lt;# elements&gt; x &lt;elementtype&gt;]
746</pre>
747
John Criswelle4c57cc2005-05-12 16:52:32 +0000748<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000749be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000750
Chris Lattner7faa8832002-04-14 06:13:44 +0000751<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000752<table class="layout">
753 <tr class="layout">
754 <td class="left">
755 <tt>[40 x int ]</tt><br/>
756 <tt>[41 x int ]</tt><br/>
757 <tt>[40 x uint]</tt><br/>
758 </td>
759 <td class="left">
760 Array of 40 integer values.<br/>
761 Array of 41 integer values.<br/>
762 Array of 40 unsigned integer values.<br/>
763 </td>
764 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000765</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000766<p>Here are some examples of multidimensional arrays:</p>
767<table class="layout">
768 <tr class="layout">
769 <td class="left">
770 <tt>[3 x [4 x int]]</tt><br/>
771 <tt>[12 x [10 x float]]</tt><br/>
772 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
773 </td>
774 <td class="left">
John Criswellc1f786c2005-05-13 22:25:59 +0000775 3x4 array of integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000776 12x10 array of single precision floating point values.<br/>
777 2x3x4 array of unsigned integer values.<br/>
778 </td>
779 </tr>
780</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000781
John Criswell0ec250c2005-10-24 16:17:18 +0000782<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
783length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000784LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
785As a special case, however, zero length arrays are recognized to be variable
786length. This allows implementation of 'pascal style arrays' with the LLVM
787type "{ int, [0 x float]}", for example.</p>
788
Misha Brukman9d0919f2003-11-08 01:05:38 +0000789</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000790
Chris Lattner00950542001-06-06 20:29:01 +0000791<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000792<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000793<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000794<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000795<p>The function type can be thought of as a function signature. It
796consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000797Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000798(which are structures of pointers to functions), for indirect function
799calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000800<p>
801The return type of a function type cannot be an aggregate type.
802</p>
Chris Lattner00950542001-06-06 20:29:01 +0000803<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000804<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000805<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +0000806specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000807which indicates that the function takes a variable number of arguments.
808Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000809 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000810<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000811<table class="layout">
812 <tr class="layout">
813 <td class="left">
814 <tt>int (int)</tt> <br/>
815 <tt>float (int, int *) *</tt><br/>
816 <tt>int (sbyte *, ...)</tt><br/>
817 </td>
818 <td class="left">
819 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
820 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukmanc24b7582004-08-12 20:16:08 +0000821 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
Reid Spencerd3f876c2004-11-01 08:19:36 +0000822 returning <tt>float</tt>.<br/>
823 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
824 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
825 the signature for <tt>printf</tt> in LLVM.<br/>
826 </td>
827 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000828</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000829
Misha Brukman9d0919f2003-11-08 01:05:38 +0000830</div>
Chris Lattner00950542001-06-06 20:29:01 +0000831<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000832<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000833<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000834<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000835<p>The structure type is used to represent a collection of data members
836together in memory. The packing of the field types is defined to match
837the ABI of the underlying processor. The elements of a structure may
838be any type that has a size.</p>
839<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
840and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
841field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
842instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000843<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000844<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000845<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000846<table class="layout">
847 <tr class="layout">
848 <td class="left">
849 <tt>{ int, int, int }</tt><br/>
850 <tt>{ float, int (int) * }</tt><br/>
851 </td>
852 <td class="left">
853 a triple of three <tt>int</tt> values<br/>
854 A pair, where the first element is a <tt>float</tt> and the second element
855 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
856 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
857 </td>
858 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000859</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000860</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000861
Chris Lattner00950542001-06-06 20:29:01 +0000862<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000863<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000864<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000865<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000866<p>As in many languages, the pointer type represents a pointer or
867reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000868<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000869<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000870<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000871<table class="layout">
872 <tr class="layout">
873 <td class="left">
874 <tt>[4x int]*</tt><br/>
875 <tt>int (int *) *</tt><br/>
876 </td>
877 <td class="left">
878 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
879 four <tt>int</tt> values<br/>
880 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattnera977c482005-02-19 02:22:14 +0000881 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerd3f876c2004-11-01 08:19:36 +0000882 <tt>int</tt>.<br/>
883 </td>
884 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000885</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000886</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000887
Chris Lattnera58561b2004-08-12 19:12:28 +0000888<!-- _______________________________________________________________________ -->
889<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000890<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000891
Chris Lattnera58561b2004-08-12 19:12:28 +0000892<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000893
Chris Lattnera58561b2004-08-12 19:12:28 +0000894<p>A packed type is a simple derived type that represents a vector
895of elements. Packed types are used when multiple primitive data
896are operated in parallel using a single instruction (SIMD).
897A packed type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000898elements) and an underlying primitive data type. Vectors must have a power
899of two length (1, 2, 4, 8, 16 ...). Packed types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000900considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000901
Chris Lattnera58561b2004-08-12 19:12:28 +0000902<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000903
904<pre>
905 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
906</pre>
907
John Criswellc1f786c2005-05-13 22:25:59 +0000908<p>The number of elements is a constant integer value; elementtype may
Chris Lattnera58561b2004-08-12 19:12:28 +0000909be any integral or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000910
Chris Lattnera58561b2004-08-12 19:12:28 +0000911<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000912
Reid Spencerd3f876c2004-11-01 08:19:36 +0000913<table class="layout">
914 <tr class="layout">
915 <td class="left">
916 <tt>&lt;4 x int&gt;</tt><br/>
917 <tt>&lt;8 x float&gt;</tt><br/>
918 <tt>&lt;2 x uint&gt;</tt><br/>
919 </td>
920 <td class="left">
921 Packed vector of 4 integer values.<br/>
922 Packed vector of 8 floating-point values.<br/>
923 Packed vector of 2 unsigned integer values.<br/>
924 </td>
925 </tr>
926</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000927</div>
928
Chris Lattner69c11bb2005-04-25 17:34:15 +0000929<!-- _______________________________________________________________________ -->
930<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
931<div class="doc_text">
932
933<h5>Overview:</h5>
934
935<p>Opaque types are used to represent unknown types in the system. This
936corresponds (for example) to the C notion of a foward declared structure type.
937In LLVM, opaque types can eventually be resolved to any type (not just a
938structure type).</p>
939
940<h5>Syntax:</h5>
941
942<pre>
943 opaque
944</pre>
945
946<h5>Examples:</h5>
947
948<table class="layout">
949 <tr class="layout">
950 <td class="left">
951 <tt>opaque</tt>
952 </td>
953 <td class="left">
954 An opaque type.<br/>
955 </td>
956 </tr>
957</table>
958</div>
959
960
Chris Lattnerc3f59762004-12-09 17:30:23 +0000961<!-- *********************************************************************** -->
962<div class="doc_section"> <a name="constants">Constants</a> </div>
963<!-- *********************************************************************** -->
964
965<div class="doc_text">
966
967<p>LLVM has several different basic types of constants. This section describes
968them all and their syntax.</p>
969
970</div>
971
972<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000973<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000974
975<div class="doc_text">
976
977<dl>
978 <dt><b>Boolean constants</b></dt>
979
980 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
981 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
982 </dd>
983
984 <dt><b>Integer constants</b></dt>
985
Reid Spencercc16dc32004-12-09 18:02:53 +0000986 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000987 href="#t_integer">integer</a> type. Negative numbers may be used with signed
988 integer types.
989 </dd>
990
991 <dt><b>Floating point constants</b></dt>
992
993 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
994 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000995 notation (see below). Floating point constants must have a <a
996 href="#t_floating">floating point</a> type. </dd>
997
998 <dt><b>Null pointer constants</b></dt>
999
John Criswell9e2485c2004-12-10 15:51:16 +00001000 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +00001001 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1002
1003</dl>
1004
John Criswell9e2485c2004-12-10 15:51:16 +00001005<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +00001006of floating point constants. For example, the form '<tt>double
10070x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
10084.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +00001009(and the only time that they are generated by the disassembler) is when a
1010floating point constant must be emitted but it cannot be represented as a
1011decimal floating point number. For example, NaN's, infinities, and other
1012special values are represented in their IEEE hexadecimal format so that
1013assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001014
1015</div>
1016
1017<!-- ======================================================================= -->
1018<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1019</div>
1020
1021<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001022<p>Aggregate constants arise from aggregation of simple constants
1023and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001024
1025<dl>
1026 <dt><b>Structure constants</b></dt>
1027
1028 <dd>Structure constants are represented with notation similar to structure
1029 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001030 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
1031 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
1032 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +00001033 types of elements must match those specified by the type.
1034 </dd>
1035
1036 <dt><b>Array constants</b></dt>
1037
1038 <dd>Array constants are represented with notation similar to array type
1039 definitions (a comma separated list of elements, surrounded by square brackets
John Criswell9e2485c2004-12-10 15:51:16 +00001040 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +00001041 constants must have <a href="#t_array">array type</a>, and the number and
1042 types of elements must match those specified by the type.
1043 </dd>
1044
1045 <dt><b>Packed constants</b></dt>
1046
1047 <dd>Packed constants are represented with notation similar to packed type
1048 definitions (a comma separated list of elements, surrounded by
John Criswell9e2485c2004-12-10 15:51:16 +00001049 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; int 42,
Chris Lattnerc3f59762004-12-09 17:30:23 +00001050 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
1051 href="#t_packed">packed type</a>, and the number and types of elements must
1052 match those specified by the type.
1053 </dd>
1054
1055 <dt><b>Zero initialization</b></dt>
1056
1057 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1058 value to zero of <em>any</em> type, including scalar and aggregate types.
1059 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +00001060 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +00001061 initializers.
1062 </dd>
1063</dl>
1064
1065</div>
1066
1067<!-- ======================================================================= -->
1068<div class="doc_subsection">
1069 <a name="globalconstants">Global Variable and Function Addresses</a>
1070</div>
1071
1072<div class="doc_text">
1073
1074<p>The addresses of <a href="#globalvars">global variables</a> and <a
1075href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +00001076constants. These constants are explicitly referenced when the <a
1077href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +00001078href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1079file:</p>
1080
1081<pre>
1082 %X = global int 17
1083 %Y = global int 42
1084 %Z = global [2 x int*] [ int* %X, int* %Y ]
1085</pre>
1086
1087</div>
1088
1089<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +00001090<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001091<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +00001092 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +00001093 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +00001094 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001095
Reid Spencer2dc45b82004-12-09 18:13:12 +00001096 <p>Undefined values indicate to the compiler that the program is well defined
1097 no matter what value is used, giving the compiler more freedom to optimize.
1098 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001099</div>
1100
1101<!-- ======================================================================= -->
1102<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1103</div>
1104
1105<div class="doc_text">
1106
1107<p>Constant expressions are used to allow expressions involving other constants
1108to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +00001109href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +00001110that does not have side effects (e.g. load and call are not supported). The
1111following is the syntax for constant expressions:</p>
1112
1113<dl>
1114 <dt><b><tt>cast ( CST to TYPE )</tt></b></dt>
1115
1116 <dd>Cast a constant to another type.</dd>
1117
1118 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1119
1120 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1121 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1122 instruction, the index list may have zero or more indexes, which are required
1123 to make sense for the type of "CSTPTR".</dd>
1124
Robert Bocchino9fbe1452006-01-10 19:31:34 +00001125 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1126
1127 <dd>Perform the <a href="#i_select">select operation</a> on
1128 constants.
1129
1130 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1131
1132 <dd>Perform the <a href="#i_extractelement">extractelement
1133 operation</a> on constants.
1134
Robert Bocchino05ccd702006-01-15 20:48:27 +00001135 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1136
1137 <dd>Perform the <a href="#i_insertelement">insertelement
1138 operation</a> on constants.
1139
Chris Lattnerc1989542006-04-08 00:13:41 +00001140
1141 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1142
1143 <dd>Perform the <a href="#i_shufflevector">shufflevector
1144 operation</a> on constants.
1145
Chris Lattnerc3f59762004-12-09 17:30:23 +00001146 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1147
Reid Spencer2dc45b82004-12-09 18:13:12 +00001148 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1149 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +00001150 binary</a> operations. The constraints on operands are the same as those for
1151 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +00001152 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001153</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001154</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00001155
Chris Lattner00950542001-06-06 20:29:01 +00001156<!-- *********************************************************************** -->
Chris Lattnere87d6532006-01-25 23:47:57 +00001157<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1158<!-- *********************************************************************** -->
1159
1160<!-- ======================================================================= -->
1161<div class="doc_subsection">
1162<a name="inlineasm">Inline Assembler Expressions</a>
1163</div>
1164
1165<div class="doc_text">
1166
1167<p>
1168LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1169Module-Level Inline Assembly</a>) through the use of a special value. This
1170value represents the inline assembler as a string (containing the instructions
1171to emit), a list of operand constraints (stored as a string), and a flag that
1172indicates whether or not the inline asm expression has side effects. An example
1173inline assembler expression is:
1174</p>
1175
1176<pre>
1177 int(int) asm "bswap $0", "=r,r"
1178</pre>
1179
1180<p>
1181Inline assembler expressions may <b>only</b> be used as the callee operand of
1182a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1183</p>
1184
1185<pre>
1186 %X = call int asm "<a href="#i_bswap">bswap</a> $0", "=r,r"(int %Y)
1187</pre>
1188
1189<p>
1190Inline asms with side effects not visible in the constraint list must be marked
1191as having side effects. This is done through the use of the
1192'<tt>sideeffect</tt>' keyword, like so:
1193</p>
1194
1195<pre>
1196 call void asm sideeffect "eieio", ""()
1197</pre>
1198
1199<p>TODO: The format of the asm and constraints string still need to be
1200documented here. Constraints on what can be done (e.g. duplication, moving, etc
1201need to be documented).
1202</p>
1203
1204</div>
1205
1206<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00001207<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1208<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00001209
Misha Brukman9d0919f2003-11-08 01:05:38 +00001210<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001211
Chris Lattner261efe92003-11-25 01:02:51 +00001212<p>The LLVM instruction set consists of several different
1213classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +00001214instructions</a>, <a href="#binaryops">binary instructions</a>,
1215<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +00001216 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1217instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001218
Misha Brukman9d0919f2003-11-08 01:05:38 +00001219</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001220
Chris Lattner00950542001-06-06 20:29:01 +00001221<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001222<div class="doc_subsection"> <a name="terminators">Terminator
1223Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001224
Misha Brukman9d0919f2003-11-08 01:05:38 +00001225<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001226
Chris Lattner261efe92003-11-25 01:02:51 +00001227<p>As mentioned <a href="#functionstructure">previously</a>, every
1228basic block in a program ends with a "Terminator" instruction, which
1229indicates which block should be executed after the current block is
1230finished. These terminator instructions typically yield a '<tt>void</tt>'
1231value: they produce control flow, not values (the one exception being
1232the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +00001233<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +00001234 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1235instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +00001236the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1237 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1238 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001239
Misha Brukman9d0919f2003-11-08 01:05:38 +00001240</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001241
Chris Lattner00950542001-06-06 20:29:01 +00001242<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001243<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1244Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001245<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001246<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001247<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 +00001248 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001249</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001250<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001251<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +00001252value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +00001253<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +00001254returns a value and then causes control flow, and one that just causes
1255control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001256<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001257<p>The '<tt>ret</tt>' instruction may return any '<a
1258 href="#t_firstclass">first class</a>' type. Notice that a function is
1259not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1260instruction inside of the function that returns a value that does not
1261match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001262<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001263<p>When the '<tt>ret</tt>' instruction is executed, control flow
1264returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +00001265 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +00001266the instruction after the call. If the caller was an "<a
1267 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +00001268at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +00001269returns a value, that value shall set the call or invoke instruction's
1270return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001271<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001272<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001273 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001274</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001275</div>
Chris Lattner00950542001-06-06 20:29:01 +00001276<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001277<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001278<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001279<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001280<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 +00001281</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001282<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001283<p>The '<tt>br</tt>' instruction is used to cause control flow to
1284transfer to a different basic block in the current function. There are
1285two forms of this instruction, corresponding to a conditional branch
1286and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001287<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001288<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1289single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1290unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1291value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001292<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001293<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1294argument is evaluated. If the value is <tt>true</tt>, control flows
1295to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1296control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001297<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001298<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
1299 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 +00001300</div>
Chris Lattner00950542001-06-06 20:29:01 +00001301<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001302<div class="doc_subsubsection">
1303 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1304</div>
1305
Misha Brukman9d0919f2003-11-08 01:05:38 +00001306<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001307<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001308
1309<pre>
1310 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1311</pre>
1312
Chris Lattner00950542001-06-06 20:29:01 +00001313<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001314
1315<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1316several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +00001317instruction, allowing a branch to occur to one of many possible
1318destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001319
1320
Chris Lattner00950542001-06-06 20:29:01 +00001321<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001322
1323<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1324comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1325an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1326table is not allowed to contain duplicate constant entries.</p>
1327
Chris Lattner00950542001-06-06 20:29:01 +00001328<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001329
Chris Lattner261efe92003-11-25 01:02:51 +00001330<p>The <tt>switch</tt> instruction specifies a table of values and
1331destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +00001332table is searched for the given value. If the value is found, control flow is
1333transfered to the corresponding destination; otherwise, control flow is
1334transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001335
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001336<h5>Implementation:</h5>
1337
1338<p>Depending on properties of the target machine and the particular
1339<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +00001340ways. For example, it could be generated as a series of chained conditional
1341branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001342
1343<h5>Example:</h5>
1344
1345<pre>
1346 <i>; Emulate a conditional br instruction</i>
1347 %Val = <a href="#i_cast">cast</a> bool %value to int
1348 switch int %Val, label %truedest [int 0, label %falsedest ]
1349
1350 <i>; Emulate an unconditional br instruction</i>
1351 switch uint 0, label %dest [ ]
1352
1353 <i>; Implement a jump table:</i>
1354 switch uint %val, label %otherwise [ uint 0, label %onzero
1355 uint 1, label %onone
1356 uint 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00001357</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001358</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001359
Chris Lattner00950542001-06-06 20:29:01 +00001360<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001361<div class="doc_subsubsection">
1362 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1363</div>
1364
Misha Brukman9d0919f2003-11-08 01:05:38 +00001365<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001366
Chris Lattner00950542001-06-06 20:29:01 +00001367<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001368
1369<pre>
1370 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)
1371 to label &lt;normal label&gt; except label &lt;exception label&gt;
1372</pre>
1373
Chris Lattner6536cfe2002-05-06 22:08:29 +00001374<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001375
1376<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1377function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +00001378'<tt>normal</tt>' label or the
1379'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001380"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1381"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +00001382href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1383continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001384
Chris Lattner00950542001-06-06 20:29:01 +00001385<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001386
Misha Brukman9d0919f2003-11-08 01:05:38 +00001387<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001388
Chris Lattner00950542001-06-06 20:29:01 +00001389<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001390 <li>
John Criswellc1f786c2005-05-13 22:25:59 +00001391 The optional "cconv" marker indicates which <a href="callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001392 convention</a> the call should use. If none is specified, the call defaults
1393 to using C calling conventions.
1394 </li>
1395 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1396 function value being invoked. In most cases, this is a direct function
1397 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1398 an arbitrary pointer to function value.
1399 </li>
1400
1401 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1402 function to be invoked. </li>
1403
1404 <li>'<tt>function args</tt>': argument list whose types match the function
1405 signature argument types. If the function signature indicates the function
1406 accepts a variable number of arguments, the extra arguments can be
1407 specified. </li>
1408
1409 <li>'<tt>normal label</tt>': the label reached when the called function
1410 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1411
1412 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1413 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1414
Chris Lattner00950542001-06-06 20:29:01 +00001415</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001416
Chris Lattner00950542001-06-06 20:29:01 +00001417<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001418
Misha Brukman9d0919f2003-11-08 01:05:38 +00001419<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001420href="#i_call">call</a></tt>' instruction in most regards. The primary
1421difference is that it establishes an association with a label, which is used by
1422the runtime library to unwind the stack.</p>
1423
1424<p>This instruction is used in languages with destructors to ensure that proper
1425cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1426exception. Additionally, this is important for implementation of
1427'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1428
Chris Lattner00950542001-06-06 20:29:01 +00001429<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001430<pre>
1431 %retval = invoke int %Test(int 15) to label %Continue
1432 except label %TestCleanup <i>; {int}:retval set</i>
1433 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
1434 except label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00001435</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001436</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001437
1438
Chris Lattner27f71f22003-09-03 00:41:47 +00001439<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00001440
Chris Lattner261efe92003-11-25 01:02:51 +00001441<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1442Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +00001443
Misha Brukman9d0919f2003-11-08 01:05:38 +00001444<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +00001445
Chris Lattner27f71f22003-09-03 00:41:47 +00001446<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001447<pre>
1448 unwind
1449</pre>
1450
Chris Lattner27f71f22003-09-03 00:41:47 +00001451<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001452
1453<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1454at the first callee in the dynamic call stack which used an <a
1455href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1456primarily used to implement exception handling.</p>
1457
Chris Lattner27f71f22003-09-03 00:41:47 +00001458<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001459
1460<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1461immediately halt. The dynamic call stack is then searched for the first <a
1462href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1463execution continues at the "exceptional" destination block specified by the
1464<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1465dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001466</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001467
1468<!-- _______________________________________________________________________ -->
1469
1470<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1471Instruction</a> </div>
1472
1473<div class="doc_text">
1474
1475<h5>Syntax:</h5>
1476<pre>
1477 unreachable
1478</pre>
1479
1480<h5>Overview:</h5>
1481
1482<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1483instruction is used to inform the optimizer that a particular portion of the
1484code is not reachable. This can be used to indicate that the code after a
1485no-return function cannot be reached, and other facts.</p>
1486
1487<h5>Semantics:</h5>
1488
1489<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1490</div>
1491
1492
1493
Chris Lattner00950542001-06-06 20:29:01 +00001494<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001495<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001496<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001497<p>Binary operators are used to do most of the computation in a
1498program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +00001499produce a single value. The operands might represent
Chris Lattnera58561b2004-08-12 19:12:28 +00001500multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1501The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +00001502necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001503<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001504</div>
Chris Lattner00950542001-06-06 20:29:01 +00001505<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001506<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1507Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001508<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001509<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001510<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 +00001511</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001512<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001513<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001514<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001515<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +00001516 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1517 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1518Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001519<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001520<p>The value produced is the integer or floating point sum of the two
1521operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001522<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001523<pre> &lt;result&gt; = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001524</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001525</div>
Chris Lattner00950542001-06-06 20:29:01 +00001526<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001527<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1528Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001529<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001530<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001531<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 +00001532</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001533<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001534<p>The '<tt>sub</tt>' instruction returns the difference of its two
1535operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001536<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1537instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001538<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001539<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001540 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001541values.
1542This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1543Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001544<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001545<p>The value produced is the integer or floating point difference of
1546the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001547<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001548<pre> &lt;result&gt; = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001549 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1550</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001551</div>
Chris Lattner00950542001-06-06 20:29:01 +00001552<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001553<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1554Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001555<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001556<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001557<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 +00001558</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001559<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001560<p>The '<tt>mul</tt>' instruction returns the product of its two
1561operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001562<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001563<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001564 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001565values.
1566This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1567Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001568<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001569<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +00001570two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001571<p>There is no signed vs unsigned multiplication. The appropriate
1572action is taken based on the type of the operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001573<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001574<pre> &lt;result&gt; = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001575</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001576</div>
Chris Lattner00950542001-06-06 20:29:01 +00001577<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001578<div class="doc_subsubsection"> <a name="i_div">'<tt>div</tt>'
1579Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001580<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001581<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001582<pre> &lt;result&gt; = div &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1583</pre>
1584<h5>Overview:</h5>
1585<p>The '<tt>div</tt>' instruction returns the quotient of its two
1586operands.</p>
1587<h5>Arguments:</h5>
1588<p>The two arguments to the '<tt>div</tt>' instruction must be either <a
1589 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001590values.
1591This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1592Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001593<h5>Semantics:</h5>
1594<p>The value produced is the integer or floating point quotient of the
1595two operands.</p>
1596<h5>Example:</h5>
1597<pre> &lt;result&gt; = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1598</pre>
1599</div>
1600<!-- _______________________________________________________________________ -->
1601<div class="doc_subsubsection"> <a name="i_rem">'<tt>rem</tt>'
1602Instruction</a> </div>
1603<div class="doc_text">
1604<h5>Syntax:</h5>
1605<pre> &lt;result&gt; = rem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1606</pre>
1607<h5>Overview:</h5>
1608<p>The '<tt>rem</tt>' instruction returns the remainder from the
1609division of its two operands.</p>
1610<h5>Arguments:</h5>
1611<p>The two arguments to the '<tt>rem</tt>' instruction must be either <a
1612 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001613values.
1614This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1615Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001616<h5>Semantics:</h5>
1617<p>This returns the <i>remainder</i> of a division (where the result
1618has the same sign as the divisor), not the <i>modulus</i> (where the
1619result has the same sign as the dividend) of a value. For more
John Criswell0ec250c2005-10-24 16:17:18 +00001620information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +00001621 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1622Math Forum</a>.</p>
1623<h5>Example:</h5>
1624<pre> &lt;result&gt; = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1625</pre>
Robert Bocchino7b81c752006-02-17 21:18:08 +00001626
Chris Lattner261efe92003-11-25 01:02:51 +00001627</div>
1628<!-- _______________________________________________________________________ -->
1629<div class="doc_subsubsection"> <a name="i_setcc">'<tt>set<i>cc</i></tt>'
1630Instructions</a> </div>
1631<div class="doc_text">
1632<h5>Syntax:</h5>
1633<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 +00001634 &lt;result&gt; = setne &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1635 &lt;result&gt; = setlt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1636 &lt;result&gt; = setgt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1637 &lt;result&gt; = setle &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1638 &lt;result&gt; = setge &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1639</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001640<h5>Overview:</h5>
1641<p>The '<tt>set<i>cc</i></tt>' family of instructions returns a boolean
1642value based on a comparison of their two operands.</p>
1643<h5>Arguments:</h5>
1644<p>The two arguments to the '<tt>set<i>cc</i></tt>' instructions must
1645be of <a href="#t_firstclass">first class</a> type (it is not possible
1646to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1647or '<tt>void</tt>' values, etc...). Both arguments must have identical
1648types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001649<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001650<p>The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1651value if both operands are equal.<br>
1652The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1653value if both operands are unequal.<br>
1654The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1655value if the first operand is less than the second operand.<br>
1656The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1657value if the first operand is greater than the second operand.<br>
1658The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1659value if the first operand is less than or equal to the second operand.<br>
1660The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1661value if the first operand is greater than or equal to the second
1662operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001663<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001664<pre> &lt;result&gt; = seteq int 4, 5 <i>; yields {bool}:result = false</i>
Chris Lattner00950542001-06-06 20:29:01 +00001665 &lt;result&gt; = setne float 4, 5 <i>; yields {bool}:result = true</i>
1666 &lt;result&gt; = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1667 &lt;result&gt; = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1668 &lt;result&gt; = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1669 &lt;result&gt; = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1670</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001671</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +00001672
Chris Lattner00950542001-06-06 20:29:01 +00001673<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001674<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1675Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001676<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001677<p>Bitwise binary operators are used to do various forms of
1678bit-twiddling in a program. They are generally very efficient
John Criswell9e2485c2004-12-10 15:51:16 +00001679instructions and can commonly be strength reduced from other
Chris Lattner261efe92003-11-25 01:02:51 +00001680instructions. They require two operands, execute an operation on them,
1681and produce a single value. The resulting value of the bitwise binary
1682operators is always the same type as its first operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001683</div>
Chris Lattner00950542001-06-06 20:29:01 +00001684<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001685<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1686Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001687<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001688<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001689<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 +00001690</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001691<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001692<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1693its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001694<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001695<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001696 href="#t_integral">integral</a> values. Both arguments must have
1697identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001698<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001699<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001700<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001701<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +00001702<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001703 <tbody>
1704 <tr>
1705 <td>In0</td>
1706 <td>In1</td>
1707 <td>Out</td>
1708 </tr>
1709 <tr>
1710 <td>0</td>
1711 <td>0</td>
1712 <td>0</td>
1713 </tr>
1714 <tr>
1715 <td>0</td>
1716 <td>1</td>
1717 <td>0</td>
1718 </tr>
1719 <tr>
1720 <td>1</td>
1721 <td>0</td>
1722 <td>0</td>
1723 </tr>
1724 <tr>
1725 <td>1</td>
1726 <td>1</td>
1727 <td>1</td>
1728 </tr>
1729 </tbody>
1730</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001731</div>
Chris Lattner00950542001-06-06 20:29:01 +00001732<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001733<pre> &lt;result&gt; = and int 4, %var <i>; yields {int}:result = 4 &amp; %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001734 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1735 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1736</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001737</div>
Chris Lattner00950542001-06-06 20:29:01 +00001738<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001739<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001740<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001741<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001742<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 +00001743</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001744<h5>Overview:</h5>
1745<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1746or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001747<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001748<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001749 href="#t_integral">integral</a> values. Both arguments must have
1750identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001751<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001752<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001753<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001754<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001755<table border="1" cellspacing="0" cellpadding="4">
1756 <tbody>
1757 <tr>
1758 <td>In0</td>
1759 <td>In1</td>
1760 <td>Out</td>
1761 </tr>
1762 <tr>
1763 <td>0</td>
1764 <td>0</td>
1765 <td>0</td>
1766 </tr>
1767 <tr>
1768 <td>0</td>
1769 <td>1</td>
1770 <td>1</td>
1771 </tr>
1772 <tr>
1773 <td>1</td>
1774 <td>0</td>
1775 <td>1</td>
1776 </tr>
1777 <tr>
1778 <td>1</td>
1779 <td>1</td>
1780 <td>1</td>
1781 </tr>
1782 </tbody>
1783</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001784</div>
Chris Lattner00950542001-06-06 20:29:01 +00001785<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001786<pre> &lt;result&gt; = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001787 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1788 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1789</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001790</div>
Chris Lattner00950542001-06-06 20:29:01 +00001791<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001792<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1793Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001794<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001795<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001796<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 +00001797</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001798<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001799<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1800or of its two operands. The <tt>xor</tt> is used to implement the
1801"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001802<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001803<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001804 href="#t_integral">integral</a> values. Both arguments must have
1805identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001806<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001807<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001808<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001809<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001810<table border="1" cellspacing="0" cellpadding="4">
1811 <tbody>
1812 <tr>
1813 <td>In0</td>
1814 <td>In1</td>
1815 <td>Out</td>
1816 </tr>
1817 <tr>
1818 <td>0</td>
1819 <td>0</td>
1820 <td>0</td>
1821 </tr>
1822 <tr>
1823 <td>0</td>
1824 <td>1</td>
1825 <td>1</td>
1826 </tr>
1827 <tr>
1828 <td>1</td>
1829 <td>0</td>
1830 <td>1</td>
1831 </tr>
1832 <tr>
1833 <td>1</td>
1834 <td>1</td>
1835 <td>0</td>
1836 </tr>
1837 </tbody>
1838</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001839</div>
Chris Lattner261efe92003-11-25 01:02:51 +00001840<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +00001841<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001842<pre> &lt;result&gt; = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001843 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
1844 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner27f71f22003-09-03 00:41:47 +00001845 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00001846</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001847</div>
Chris Lattner00950542001-06-06 20:29:01 +00001848<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001849<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
1850Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001851<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001852<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001853<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 +00001854</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001855<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001856<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
1857the left a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001858<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001859<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001860 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1861type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001862<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001863<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001864<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001865<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 +00001866 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1867 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1868</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001869</div>
Chris Lattner00950542001-06-06 20:29:01 +00001870<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001871<div class="doc_subsubsection"> <a name="i_shr">'<tt>shr</tt>'
1872Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001873<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001874<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001875<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 +00001876</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001877<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001878<p>The '<tt>shr</tt>' instruction returns the first operand shifted to
1879the right a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001880<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001881<p>The first argument to the '<tt>shr</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00001882 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1883type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001884<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001885<p>If the first argument is a <a href="#t_signed">signed</a> type, the
1886most significant bit is duplicated in the newly free'd bit positions.
1887If the first argument is unsigned, zero bits shall fill the empty
1888positions.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001889<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001890<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 +00001891 &lt;result&gt; = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
Chris Lattner00950542001-06-06 20:29:01 +00001892 &lt;result&gt; = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +00001893 &lt;result&gt; = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1894 &lt;result&gt; = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner00950542001-06-06 20:29:01 +00001895</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001896</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00001897
Chris Lattner00950542001-06-06 20:29:01 +00001898<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00001899<div class="doc_subsection">
Chris Lattner3df241e2006-04-08 23:07:04 +00001900 <a name="vectorops">Vector Operations</a>
1901</div>
1902
1903<div class="doc_text">
1904
1905<p>LLVM supports several instructions to represent vector operations in a
1906target-independent manner. This instructions cover the element-access and
1907vector-specific operations needed to process vectors effectively. While LLVM
1908does directly support these vector operations, many sophisticated algorithms
1909will want to use target-specific intrinsics to take full advantage of a specific
1910target.</p>
1911
1912</div>
1913
1914<!-- _______________________________________________________________________ -->
1915<div class="doc_subsubsection">
1916 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
1917</div>
1918
1919<div class="doc_text">
1920
1921<h5>Syntax:</h5>
1922
1923<pre>
1924 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, uint &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
1925</pre>
1926
1927<h5>Overview:</h5>
1928
1929<p>
1930The '<tt>extractelement</tt>' instruction extracts a single scalar
1931element from a packed vector at a specified index.
1932</p>
1933
1934
1935<h5>Arguments:</h5>
1936
1937<p>
1938The first operand of an '<tt>extractelement</tt>' instruction is a
1939value of <a href="#t_packed">packed</a> type. The second operand is
1940an index indicating the position from which to extract the element.
1941The index may be a variable.</p>
1942
1943<h5>Semantics:</h5>
1944
1945<p>
1946The result is a scalar of the same type as the element type of
1947<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
1948<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
1949results are undefined.
1950</p>
1951
1952<h5>Example:</h5>
1953
1954<pre>
1955 %result = extractelement &lt;4 x int&gt; %vec, uint 0 <i>; yields int</i>
1956</pre>
1957</div>
1958
1959
1960<!-- _______________________________________________________________________ -->
1961<div class="doc_subsubsection">
1962 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
1963</div>
1964
1965<div class="doc_text">
1966
1967<h5>Syntax:</h5>
1968
1969<pre>
1970 &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>
1971</pre>
1972
1973<h5>Overview:</h5>
1974
1975<p>
1976The '<tt>insertelement</tt>' instruction inserts a scalar
1977element into a packed vector at a specified index.
1978</p>
1979
1980
1981<h5>Arguments:</h5>
1982
1983<p>
1984The first operand of an '<tt>insertelement</tt>' instruction is a
1985value of <a href="#t_packed">packed</a> type. The second operand is a
1986scalar value whose type must equal the element type of the first
1987operand. The third operand is an index indicating the position at
1988which to insert the value. The index may be a variable.</p>
1989
1990<h5>Semantics:</h5>
1991
1992<p>
1993The result is a packed vector of the same type as <tt>val</tt>. Its
1994element values are those of <tt>val</tt> except at position
1995<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
1996exceeds the length of <tt>val</tt>, the results are undefined.
1997</p>
1998
1999<h5>Example:</h5>
2000
2001<pre>
2002 %result = insertelement &lt;4 x int&gt; %vec, int 1, uint 0 <i>; yields &lt;4 x int&gt;</i>
2003</pre>
2004</div>
2005
2006<!-- _______________________________________________________________________ -->
2007<div class="doc_subsubsection">
2008 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2009</div>
2010
2011<div class="doc_text">
2012
2013<h5>Syntax:</h5>
2014
2015<pre>
2016 &lt;result&gt; = shufflevector &lt;n x &lt;ty&gt;&gt; &lt;v1&gt;, &lt;n x &lt;ty&gt;&gt; &lt;v2&gt;, &lt;n x uint&gt; &lt;mask&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2017</pre>
2018
2019<h5>Overview:</h5>
2020
2021<p>
2022The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2023from two input vectors, returning a vector of the same type.
2024</p>
2025
2026<h5>Arguments:</h5>
2027
2028<p>
2029The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2030with types that match each other and types that match the result of the
2031instruction. The third argument is a shuffle mask, which has the same number
2032of elements as the other vector type, but whose element type is always 'uint'.
2033</p>
2034
2035<p>
2036The shuffle mask operand is required to be a constant vector with either
2037constant integer or undef values.
2038</p>
2039
2040<h5>Semantics:</h5>
2041
2042<p>
2043The elements of the two input vectors are numbered from left to right across
2044both of the vectors. The shuffle mask operand specifies, for each element of
2045the result vector, which element of the two input registers the result element
2046gets. The element selector may be undef (meaning "don't care") and the second
2047operand may be undef if performing a shuffle from only one vector.
2048</p>
2049
2050<h5>Example:</h5>
2051
2052<pre>
2053 %result = shufflevector &lt;4 x int&gt; %v1, &lt;4 x int&gt; %v2,
2054 &lt;4 x uint&gt; &lt;uint 0, uint 4, uint 1, uint 5&gt; <i>; yields &lt;4 x int&gt;</i>
2055 %result = shufflevector &lt;4 x int&gt; %v1, &lt;4 x int&gt; undef,
2056 &lt;4 x uint&gt; &lt;uint 0, uint 1, uint 2, uint 3&gt; <i>; yields &lt;4 x int&gt;</i> - Identity shuffle.
2057</pre>
2058</div>
2059
Chris Lattner3df241e2006-04-08 23:07:04 +00002060<!-- ======================================================================= -->
2061<div class="doc_subsection">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002062 <a name="memoryops">Memory Access Operations</a>
2063</div>
2064
Misha Brukman9d0919f2003-11-08 01:05:38 +00002065<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002066
Chris Lattner261efe92003-11-25 01:02:51 +00002067<p>A key design point of an SSA-based representation is how it
2068represents memory. In LLVM, no memory locations are in SSA form, which
2069makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +00002070allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002071
Misha Brukman9d0919f2003-11-08 01:05:38 +00002072</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002073
Chris Lattner00950542001-06-06 20:29:01 +00002074<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002075<div class="doc_subsubsection">
2076 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2077</div>
2078
Misha Brukman9d0919f2003-11-08 01:05:38 +00002079<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002080
Chris Lattner00950542001-06-06 20:29:01 +00002081<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002082
2083<pre>
2084 &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 +00002085</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002086
Chris Lattner00950542001-06-06 20:29:01 +00002087<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002088
Chris Lattner261efe92003-11-25 01:02:51 +00002089<p>The '<tt>malloc</tt>' instruction allocates memory from the system
2090heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002091
Chris Lattner00950542001-06-06 20:29:01 +00002092<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002093
2094<p>The '<tt>malloc</tt>' instruction allocates
2095<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +00002096bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00002097appropriate type to the program. If "NumElements" is specified, it is the
2098number of elements allocated. If an alignment is specified, the value result
2099of the allocation is guaranteed to be aligned to at least that boundary. If
2100not specified, or if zero, the target can choose to align the allocation on any
2101convenient boundary.</p>
2102
Misha Brukman9d0919f2003-11-08 01:05:38 +00002103<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002104
Chris Lattner00950542001-06-06 20:29:01 +00002105<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002106
Chris Lattner261efe92003-11-25 01:02:51 +00002107<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2108a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002109
Chris Lattner2cbdc452005-11-06 08:02:57 +00002110<h5>Example:</h5>
2111
2112<pre>
2113 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
2114
2115 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002116 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
2117 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002118 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
2119 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +00002120</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002121</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002122
Chris Lattner00950542001-06-06 20:29:01 +00002123<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002124<div class="doc_subsubsection">
2125 <a name="i_free">'<tt>free</tt>' Instruction</a>
2126</div>
2127
Misha Brukman9d0919f2003-11-08 01:05:38 +00002128<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002129
Chris Lattner00950542001-06-06 20:29:01 +00002130<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002131
2132<pre>
2133 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +00002134</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002135
Chris Lattner00950542001-06-06 20:29:01 +00002136<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002137
Chris Lattner261efe92003-11-25 01:02:51 +00002138<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +00002139memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002140
Chris Lattner00950542001-06-06 20:29:01 +00002141<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002142
Chris Lattner261efe92003-11-25 01:02:51 +00002143<p>'<tt>value</tt>' shall be a pointer value that points to a value
2144that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2145instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002146
Chris Lattner00950542001-06-06 20:29:01 +00002147<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002148
John Criswell9e2485c2004-12-10 15:51:16 +00002149<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +00002150after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002151
Chris Lattner00950542001-06-06 20:29:01 +00002152<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002153
2154<pre>
2155 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
Chris Lattner00950542001-06-06 20:29:01 +00002156 free [4 x ubyte]* %array
2157</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002158</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002159
Chris Lattner00950542001-06-06 20:29:01 +00002160<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002161<div class="doc_subsubsection">
2162 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2163</div>
2164
Misha Brukman9d0919f2003-11-08 01:05:38 +00002165<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002166
Chris Lattner00950542001-06-06 20:29:01 +00002167<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002168
2169<pre>
2170 &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 +00002171</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002172
Chris Lattner00950542001-06-06 20:29:01 +00002173<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002174
Chris Lattner261efe92003-11-25 01:02:51 +00002175<p>The '<tt>alloca</tt>' instruction allocates memory on the current
2176stack frame of the procedure that is live until the current function
2177returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002178
Chris Lattner00950542001-06-06 20:29:01 +00002179<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002180
John Criswell9e2485c2004-12-10 15:51:16 +00002181<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002182bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00002183appropriate type to the program. If "NumElements" is specified, it is the
2184number of elements allocated. If an alignment is specified, the value result
2185of the allocation is guaranteed to be aligned to at least that boundary. If
2186not specified, or if zero, the target can choose to align the allocation on any
2187convenient boundary.</p>
2188
Misha Brukman9d0919f2003-11-08 01:05:38 +00002189<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002190
Chris Lattner00950542001-06-06 20:29:01 +00002191<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002192
John Criswellc1f786c2005-05-13 22:25:59 +00002193<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +00002194memory is automatically released when the function returns. The '<tt>alloca</tt>'
2195instruction is commonly used to represent automatic variables that must
2196have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +00002197 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002198instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002199
Chris Lattner00950542001-06-06 20:29:01 +00002200<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002201
2202<pre>
2203 %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002204 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002205 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
2206 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00002207</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002208</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002209
Chris Lattner00950542001-06-06 20:29:01 +00002210<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002211<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2212Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002213<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00002214<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002215<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 +00002216<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002217<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002218<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002219<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +00002220address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +00002221 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +00002222marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +00002223the number or order of execution of this <tt>load</tt> with other
2224volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2225instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002226<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002227<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002228<h5>Examples:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002229<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2230 <a
2231 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002232 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2233</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002234</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002235<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002236<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2237Instruction</a> </div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002238<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002239<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 +00002240 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 +00002241</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002242<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002243<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002244<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002245<p>There are two arguments to the '<tt>store</tt>' instruction: a value
John Criswell0ec250c2005-10-24 16:17:18 +00002246to 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 +00002247operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +00002248operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +00002249optimizer is not allowed to modify the number or order of execution of
2250this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2251 href="#i_store">store</a></tt> instructions.</p>
2252<h5>Semantics:</h5>
2253<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2254at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002255<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002256<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2257 <a
2258 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002259 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2260</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002261<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002262<div class="doc_subsubsection">
2263 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2264</div>
2265
Misha Brukman9d0919f2003-11-08 01:05:38 +00002266<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +00002267<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002268<pre>
2269 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2270</pre>
2271
Chris Lattner7faa8832002-04-14 06:13:44 +00002272<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002273
2274<p>
2275The '<tt>getelementptr</tt>' instruction is used to get the address of a
2276subelement of an aggregate data structure.</p>
2277
Chris Lattner7faa8832002-04-14 06:13:44 +00002278<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002279
2280<p>This instruction takes a list of integer constants that indicate what
2281elements of the aggregate object to index to. The actual types of the arguments
2282provided depend on the type of the first pointer argument. The
2283'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +00002284levels of a structure or to a specific index in an array. When indexing into a
2285structure, only <tt>uint</tt>
John Criswellc1f786c2005-05-13 22:25:59 +00002286integer constants are allowed. When indexing into an array or pointer,
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002287<tt>int</tt> and <tt>long</tt> indexes are allowed of any sign.</p>
2288
Chris Lattner261efe92003-11-25 01:02:51 +00002289<p>For example, let's consider a C code fragment and how it gets
2290compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002291
2292<pre>
2293 struct RT {
2294 char A;
2295 int B[10][20];
2296 char C;
2297 };
2298 struct ST {
2299 int X;
2300 double Y;
2301 struct RT Z;
2302 };
2303
2304 int *foo(struct ST *s) {
2305 return &amp;s[1].Z.B[5][13];
2306 }
2307</pre>
2308
Misha Brukman9d0919f2003-11-08 01:05:38 +00002309<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002310
2311<pre>
2312 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2313 %ST = type { int, double, %RT }
2314
Brian Gaeke7283e7c2004-07-02 21:08:14 +00002315 implementation
2316
2317 int* %foo(%ST* %s) {
2318 entry:
2319 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002320 ret int* %reg
2321 }
2322</pre>
2323
Chris Lattner7faa8832002-04-14 06:13:44 +00002324<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002325
2326<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +00002327on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Chris Lattnere53e5082004-06-03 22:57:15 +00002328and <a href="#t_array">array</a> types require <tt>uint</tt>, <tt>int</tt>,
2329<tt>ulong</tt>, or <tt>long</tt> values, and <a href="#t_struct">structure</a>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002330types require <tt>uint</tt> <b>constants</b>.</p>
2331
Misha Brukman9d0919f2003-11-08 01:05:38 +00002332<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002333type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2334}</tt>' type, a structure. The second index indexes into the third element of
2335the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2336sbyte }</tt>' type, another structure. The third index indexes into the second
2337element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2338array. The two dimensions of the array are subscripted into, yielding an
John Criswellfc6b8952005-05-16 16:17:45 +00002339'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002340to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2341
Chris Lattner261efe92003-11-25 01:02:51 +00002342<p>Note that it is perfectly legal to index partially through a
2343structure, returning a pointer to an inner element. Because of this,
2344the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002345
2346<pre>
Chris Lattnerd4f6b172005-03-07 22:13:59 +00002347 int* %foo(%ST* %s) {
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002348 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2349 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2350 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2351 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2352 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2353 ret int* %t5
2354 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00002355</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00002356
2357<p>Note that it is undefined to access an array out of bounds: array and
2358pointer indexes must always be within the defined bounds of the array type.
2359The one exception for this rules is zero length arrays. These arrays are
2360defined to be accessible as variable length arrays, which requires access
2361beyond the zero'th element.</p>
2362
Chris Lattner7faa8832002-04-14 06:13:44 +00002363<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +00002364
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002365<pre>
2366 <i>; yields [12 x ubyte]*:aptr</i>
2367 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2368</pre>
2369
2370</div>
Chris Lattner00950542001-06-06 20:29:01 +00002371<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00002372<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002373<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +00002374<p>The instructions in this category are the "miscellaneous"
Chris Lattner261efe92003-11-25 01:02:51 +00002375instructions, which defy better classification.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002376</div>
Chris Lattner00950542001-06-06 20:29:01 +00002377<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002378<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
2379Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002380<div class="doc_text">
Chris Lattner33ba0d92001-07-09 00:26:23 +00002381<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002382<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002383<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002384<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
2385the SSA graph representing the function.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002386<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002387<p>The type of the incoming values are specified with the first type
2388field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
2389as arguments, with one pair for each predecessor basic block of the
2390current block. Only values of <a href="#t_firstclass">first class</a>
2391type may be used as the value arguments to the PHI node. Only labels
2392may be used as the label arguments.</p>
2393<p>There must be no non-phi instructions between the start of a basic
2394block and the PHI instructions: i.e. PHI instructions must be first in
2395a basic block.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002396<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002397<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
2398value specified by the parameter, depending on which basic block we
2399came from in the last <a href="#terminators">terminator</a> instruction.</p>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002400<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002401<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 +00002402</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002403
Chris Lattner6536cfe2002-05-06 22:08:29 +00002404<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +00002405<div class="doc_subsubsection">
2406 <a name="i_cast">'<tt>cast .. to</tt>' Instruction</a>
2407</div>
2408
Misha Brukman9d0919f2003-11-08 01:05:38 +00002409<div class="doc_text">
Chris Lattnercc37aae2004-03-12 05:50:16 +00002410
Chris Lattner6536cfe2002-05-06 22:08:29 +00002411<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002412
2413<pre>
2414 &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 +00002415</pre>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002416
Chris Lattner6536cfe2002-05-06 22:08:29 +00002417<h5>Overview:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002418
2419<p>
2420The '<tt>cast</tt>' instruction is used as the primitive means to convert
2421integers to floating point, change data type sizes, and break type safety (by
2422casting pointers).
2423</p>
2424
2425
Chris Lattner6536cfe2002-05-06 22:08:29 +00002426<h5>Arguments:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002427
2428<p>
2429The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
2430class value, and a type to cast it to, which must also be a <a
2431href="#t_firstclass">first class</a> type.
2432</p>
2433
Chris Lattner6536cfe2002-05-06 22:08:29 +00002434<h5>Semantics:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002435
2436<p>
2437This instruction follows the C rules for explicit casts when determining how the
2438data being cast must change to fit in its new container.
2439</p>
2440
2441<p>
2442When casting to bool, any value that would be considered true in the context of
2443a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
2444all else are '<tt>false</tt>'.
2445</p>
2446
2447<p>
2448When extending an integral value from a type of one signness to another (for
2449example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
2450<b>source</b> value is signed, and zero-extended if the source value is
2451unsigned. <tt>bool</tt> values are always zero extended into either zero or
2452one.
2453</p>
2454
Chris Lattner33ba0d92001-07-09 00:26:23 +00002455<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002456
2457<pre>
2458 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
Chris Lattner7bae3952002-06-25 18:03:17 +00002459 %Y = cast int 123 to bool <i>; yields bool:true</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00002460</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002461</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002462
2463<!-- _______________________________________________________________________ -->
2464<div class="doc_subsubsection">
2465 <a name="i_select">'<tt>select</tt>' Instruction</a>
2466</div>
2467
2468<div class="doc_text">
2469
2470<h5>Syntax:</h5>
2471
2472<pre>
2473 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
2474</pre>
2475
2476<h5>Overview:</h5>
2477
2478<p>
2479The '<tt>select</tt>' instruction is used to choose one value based on a
2480condition, without branching.
2481</p>
2482
2483
2484<h5>Arguments:</h5>
2485
2486<p>
2487The '<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.
2488</p>
2489
2490<h5>Semantics:</h5>
2491
2492<p>
2493If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +00002494value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +00002495</p>
2496
2497<h5>Example:</h5>
2498
2499<pre>
2500 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
2501</pre>
2502</div>
2503
Robert Bocchino05ccd702006-01-15 20:48:27 +00002504
2505<!-- _______________________________________________________________________ -->
2506<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +00002507 <a name="i_call">'<tt>call</tt>' Instruction</a>
2508</div>
2509
Misha Brukman9d0919f2003-11-08 01:05:38 +00002510<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +00002511
Chris Lattner00950542001-06-06 20:29:01 +00002512<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002513<pre>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002514 &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 +00002515</pre>
2516
Chris Lattner00950542001-06-06 20:29:01 +00002517<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002518
Misha Brukman9d0919f2003-11-08 01:05:38 +00002519<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002520
Chris Lattner00950542001-06-06 20:29:01 +00002521<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002522
Misha Brukman9d0919f2003-11-08 01:05:38 +00002523<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002524
Chris Lattner6536cfe2002-05-06 22:08:29 +00002525<ol>
Chris Lattner261efe92003-11-25 01:02:51 +00002526 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002527 <p>The optional "tail" marker indicates whether the callee function accesses
2528 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +00002529 function call is eligible for tail call optimization. Note that calls may
2530 be marked "tail" even if they do not occur before a <a
2531 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +00002532 </li>
2533 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002534 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
2535 convention</a> the call should use. If none is specified, the call defaults
2536 to using C calling conventions.
2537 </li>
2538 <li>
Chris Lattner2bff5242005-05-06 05:47:36 +00002539 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
2540 being invoked. The argument types must match the types implied by this
John Criswellfc6b8952005-05-16 16:17:45 +00002541 signature. This type can be omitted if the function is not varargs and
2542 if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002543 </li>
2544 <li>
2545 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
2546 be invoked. In most cases, this is a direct function invocation, but
2547 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +00002548 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002549 </li>
2550 <li>
2551 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +00002552 function signature argument types. All arguments must be of
2553 <a href="#t_firstclass">first class</a> type. If the function signature
2554 indicates the function accepts a variable number of arguments, the extra
2555 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002556 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00002557</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00002558
Chris Lattner00950542001-06-06 20:29:01 +00002559<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002560
Chris Lattner261efe92003-11-25 01:02:51 +00002561<p>The '<tt>call</tt>' instruction is used to cause control flow to
2562transfer to a specified function, with its incoming arguments bound to
2563the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
2564instruction in the called function, control flow continues with the
2565instruction after the function call, and the return value of the
2566function is bound to the result argument. This is a simpler case of
2567the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00002568
Chris Lattner00950542001-06-06 20:29:01 +00002569<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00002570
2571<pre>
2572 %retval = call int %test(int %argc)
2573 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
2574 %X = tail call int %foo()
Chris Lattnerbad10ee2005-05-06 22:57:40 +00002575 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
Chris Lattner2bff5242005-05-06 05:47:36 +00002576</pre>
2577
Misha Brukman9d0919f2003-11-08 01:05:38 +00002578</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002579
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002580<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +00002581<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +00002582 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002583</div>
2584
Misha Brukman9d0919f2003-11-08 01:05:38 +00002585<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +00002586
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002587<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002588
2589<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002590 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00002591</pre>
2592
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002593<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002594
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002595<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +00002596the "variable argument" area of a function call. It is used to implement the
2597<tt>va_arg</tt> macro in C.</p>
2598
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002599<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002600
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002601<p>This instruction takes a <tt>va_list*</tt> value and the type of
2602the argument. It returns a value of the specified argument type and
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00002603increments the <tt>va_list</tt> to point to the next argument. Again, the
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002604actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002605
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002606<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002607
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002608<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
2609type from the specified <tt>va_list</tt> and causes the
2610<tt>va_list</tt> to point to the next argument. For more information,
2611see the variable argument handling <a href="#int_varargs">Intrinsic
2612Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002613
2614<p>It is legal for this instruction to be called in a function which does not
2615take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002616function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002617
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002618<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +00002619href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +00002620argument.</p>
2621
Chris Lattner8d1a81d2003-10-18 05:51:36 +00002622<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00002623
2624<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
2625
Misha Brukman9d0919f2003-11-08 01:05:38 +00002626</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002627
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002628<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00002629<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
2630<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002631
Misha Brukman9d0919f2003-11-08 01:05:38 +00002632<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +00002633
2634<p>LLVM supports the notion of an "intrinsic function". These functions have
John Criswellfc6b8952005-05-16 16:17:45 +00002635well known names and semantics and are required to follow certain
Chris Lattner33aec9e2004-02-12 17:01:32 +00002636restrictions. Overall, these instructions represent an extension mechanism for
2637the LLVM language that does not require changing all of the transformations in
2638LLVM to add to the language (or the bytecode reader/writer, the parser,
2639etc...).</p>
2640
John Criswellfc6b8952005-05-16 16:17:45 +00002641<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
2642prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
Chris Lattner33aec9e2004-02-12 17:01:32 +00002643this. Intrinsic functions must always be external functions: you cannot define
2644the body of intrinsic functions. Intrinsic functions may only be used in call
2645or invoke instructions: it is illegal to take the address of an intrinsic
2646function. Additionally, because intrinsic functions are part of the LLVM
2647language, it is required that they all be documented here if any are added.</p>
2648
2649
John Criswellfc6b8952005-05-16 16:17:45 +00002650<p>To learn how to add an intrinsic function, please see the <a
Chris Lattner590cff32005-05-11 03:35:57 +00002651href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +00002652</p>
2653
Misha Brukman9d0919f2003-11-08 01:05:38 +00002654</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002655
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002656<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002657<div class="doc_subsection">
2658 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
2659</div>
2660
Misha Brukman9d0919f2003-11-08 01:05:38 +00002661<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002662
Misha Brukman9d0919f2003-11-08 01:05:38 +00002663<p>Variable argument support is defined in LLVM with the <a
Chris Lattnerfb6977d2006-01-13 23:26:01 +00002664 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner261efe92003-11-25 01:02:51 +00002665intrinsic functions. These functions are related to the similarly
2666named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002667
Chris Lattner261efe92003-11-25 01:02:51 +00002668<p>All of these functions operate on arguments that use a
2669target-specific value type "<tt>va_list</tt>". The LLVM assembly
2670language reference manual does not define what this type is, so all
2671transformations should be prepared to handle intrinsics with any type
2672used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002673
Misha Brukman9d0919f2003-11-08 01:05:38 +00002674<p>This example shows how the <a href="#i_vanext"><tt>vanext</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +00002675instruction and the variable argument handling intrinsic functions are
2676used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002677
Chris Lattner33aec9e2004-02-12 17:01:32 +00002678<pre>
2679int %test(int %X, ...) {
2680 ; Initialize variable argument processing
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002681 %ap = alloca sbyte*
2682 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002683
2684 ; Read a single integer argument
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002685 %tmp = va_arg sbyte** %ap, int
Chris Lattner33aec9e2004-02-12 17:01:32 +00002686
2687 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002688 %aq = alloca sbyte*
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002689 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002690 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002691
2692 ; Stop processing of arguments.
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002693 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00002694 ret int %tmp
2695}
2696</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002697</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002698
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002699<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002700<div class="doc_subsubsection">
2701 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
2702</div>
2703
2704
Misha Brukman9d0919f2003-11-08 01:05:38 +00002705<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002706<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002707<pre> declare void %llvm.va_start(&lt;va_list&gt;* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002708<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002709<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
2710<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
2711href="#i_va_arg">va_arg</a></tt>.</p>
2712
2713<h5>Arguments:</h5>
2714
2715<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
2716
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002717<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002718
2719<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
2720macro available in C. In a target-dependent way, it initializes the
2721<tt>va_list</tt> element the argument points to, so that the next call to
2722<tt>va_arg</tt> will produce the first variable argument passed to the function.
2723Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
2724last argument of the function, the compiler can figure that out.</p>
2725
Misha Brukman9d0919f2003-11-08 01:05:38 +00002726</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002727
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002728<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002729<div class="doc_subsubsection">
2730 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
2731</div>
2732
Misha Brukman9d0919f2003-11-08 01:05:38 +00002733<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002734<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002735<pre> declare void %llvm.va_end(&lt;va_list*&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002736<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002737<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
2738which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
2739or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002740<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002741<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002742<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002743<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002744macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
2745Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
2746 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
2747with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002748</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002749
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002750<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00002751<div class="doc_subsubsection">
2752 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
2753</div>
2754
Misha Brukman9d0919f2003-11-08 01:05:38 +00002755<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00002756
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002757<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002758
2759<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002760 declare void %llvm.va_copy(&lt;va_list&gt;* &lt;destarglist&gt;,
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002761 &lt;va_list&gt;* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00002762</pre>
2763
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002764<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002765
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002766<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
2767the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002768
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002769<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002770
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002771<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00002772The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002773
Chris Lattnerd7923912004-05-23 21:06:01 +00002774
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00002775<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00002776
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00002777<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
2778available in C. In a target-dependent way, it copies the source
2779<tt>va_list</tt> element into the destination list. This intrinsic is necessary
2780because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
Chris Lattnerd7923912004-05-23 21:06:01 +00002781arbitrarily complex and require memory allocation, for example.</p>
2782
Misha Brukman9d0919f2003-11-08 01:05:38 +00002783</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00002784
Chris Lattner33aec9e2004-02-12 17:01:32 +00002785<!-- ======================================================================= -->
2786<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +00002787 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
2788</div>
2789
2790<div class="doc_text">
2791
2792<p>
2793LLVM support for <a href="GarbageCollection.html">Accurate Garbage
2794Collection</a> requires the implementation and generation of these intrinsics.
2795These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
2796stack</a>, as well as garbage collector implementations that require <a
2797href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
2798Front-ends for type-safe garbage collected languages should generate these
2799intrinsics to make use of the LLVM garbage collectors. For more details, see <a
2800href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
2801</p>
2802</div>
2803
2804<!-- _______________________________________________________________________ -->
2805<div class="doc_subsubsection">
2806 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
2807</div>
2808
2809<div class="doc_text">
2810
2811<h5>Syntax:</h5>
2812
2813<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00002814 declare void %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00002815</pre>
2816
2817<h5>Overview:</h5>
2818
John Criswell9e2485c2004-12-10 15:51:16 +00002819<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +00002820the code generator, and allows some metadata to be associated with it.</p>
2821
2822<h5>Arguments:</h5>
2823
2824<p>The first argument specifies the address of a stack object that contains the
2825root pointer. The second pointer (which must be either a constant or a global
2826value address) contains the meta-data to be associated with the root.</p>
2827
2828<h5>Semantics:</h5>
2829
2830<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
2831location. At compile-time, the code generator generates information to allow
2832the runtime to find the pointer at GC safe points.
2833</p>
2834
2835</div>
2836
2837
2838<!-- _______________________________________________________________________ -->
2839<div class="doc_subsubsection">
2840 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
2841</div>
2842
2843<div class="doc_text">
2844
2845<h5>Syntax:</h5>
2846
2847<pre>
Chris Lattner80626e92006-03-14 20:02:51 +00002848 declare sbyte* %llvm.gcread(sbyte* %ObjPtr, sbyte** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00002849</pre>
2850
2851<h5>Overview:</h5>
2852
2853<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
2854locations, allowing garbage collector implementations that require read
2855barriers.</p>
2856
2857<h5>Arguments:</h5>
2858
Chris Lattner80626e92006-03-14 20:02:51 +00002859<p>The second argument is the address to read from, which should be an address
2860allocated from the garbage collector. The first object is a pointer to the
2861start of the referenced object, if needed by the language runtime (otherwise
2862null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002863
2864<h5>Semantics:</h5>
2865
2866<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
2867instruction, but may be replaced with substantially more complex code by the
2868garbage collector runtime, as needed.</p>
2869
2870</div>
2871
2872
2873<!-- _______________________________________________________________________ -->
2874<div class="doc_subsubsection">
2875 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
2876</div>
2877
2878<div class="doc_text">
2879
2880<h5>Syntax:</h5>
2881
2882<pre>
Chris Lattner80626e92006-03-14 20:02:51 +00002883 declare void %llvm.gcwrite(sbyte* %P1, sbyte* %Obj, sbyte** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00002884</pre>
2885
2886<h5>Overview:</h5>
2887
2888<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
2889locations, allowing garbage collector implementations that require write
2890barriers (such as generational or reference counting collectors).</p>
2891
2892<h5>Arguments:</h5>
2893
Chris Lattner80626e92006-03-14 20:02:51 +00002894<p>The first argument is the reference to store, the second is the start of the
2895object to store it to, and the third is the address of the field of Obj to
2896store to. If the runtime does not require a pointer to the object, Obj may be
2897null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00002898
2899<h5>Semantics:</h5>
2900
2901<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
2902instruction, but may be replaced with substantially more complex code by the
2903garbage collector runtime, as needed.</p>
2904
2905</div>
2906
2907
2908
2909<!-- ======================================================================= -->
2910<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +00002911 <a name="int_codegen">Code Generator Intrinsics</a>
2912</div>
2913
2914<div class="doc_text">
2915<p>
2916These intrinsics are provided by LLVM to expose special features that may only
2917be implemented with code generator support.
2918</p>
2919
2920</div>
2921
2922<!-- _______________________________________________________________________ -->
2923<div class="doc_subsubsection">
2924 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
2925</div>
2926
2927<div class="doc_text">
2928
2929<h5>Syntax:</h5>
2930<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00002931 declare sbyte *%llvm.returnaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00002932</pre>
2933
2934<h5>Overview:</h5>
2935
2936<p>
2937The '<tt>llvm.returnaddress</tt>' intrinsic returns a target-specific value
2938indicating the return address of the current function or one of its callers.
2939</p>
2940
2941<h5>Arguments:</h5>
2942
2943<p>
2944The argument to this intrinsic indicates which function to return the address
2945for. Zero indicates the calling function, one indicates its caller, etc. The
2946argument is <b>required</b> to be a constant integer value.
2947</p>
2948
2949<h5>Semantics:</h5>
2950
2951<p>
2952The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
2953the return address of the specified call frame, or zero if it cannot be
2954identified. The value returned by this intrinsic is likely to be incorrect or 0
2955for arguments other than zero, so it should only be used for debugging purposes.
2956</p>
2957
2958<p>
2959Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00002960aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00002961source-language caller.
2962</p>
2963</div>
2964
2965
2966<!-- _______________________________________________________________________ -->
2967<div class="doc_subsubsection">
2968 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
2969</div>
2970
2971<div class="doc_text">
2972
2973<h5>Syntax:</h5>
2974<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00002975 declare sbyte *%llvm.frameaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00002976</pre>
2977
2978<h5>Overview:</h5>
2979
2980<p>
2981The '<tt>llvm.frameaddress</tt>' intrinsic returns the target-specific frame
2982pointer value for the specified stack frame.
2983</p>
2984
2985<h5>Arguments:</h5>
2986
2987<p>
2988The argument to this intrinsic indicates which function to return the frame
2989pointer for. Zero indicates the calling function, one indicates its caller,
2990etc. The argument is <b>required</b> to be a constant integer value.
2991</p>
2992
2993<h5>Semantics:</h5>
2994
2995<p>
2996The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
2997the frame address of the specified call frame, or zero if it cannot be
2998identified. The value returned by this intrinsic is likely to be incorrect or 0
2999for arguments other than zero, so it should only be used for debugging purposes.
3000</p>
3001
3002<p>
3003Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00003004aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00003005source-language caller.
3006</p>
3007</div>
3008
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003009<!-- _______________________________________________________________________ -->
3010<div class="doc_subsubsection">
Chris Lattner57e1f392006-01-13 02:03:13 +00003011 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
3012</div>
3013
3014<div class="doc_text">
3015
3016<h5>Syntax:</h5>
3017<pre>
3018 declare sbyte *%llvm.stacksave()
3019</pre>
3020
3021<h5>Overview:</h5>
3022
3023<p>
3024The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
3025the function stack, for use with <a href="#i_stackrestore">
3026<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
3027features like scoped automatic variable sized arrays in C99.
3028</p>
3029
3030<h5>Semantics:</h5>
3031
3032<p>
3033This intrinsic returns a opaque pointer value that can be passed to <a
3034href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
3035<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
3036<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
3037state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
3038practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
3039that were allocated after the <tt>llvm.stacksave</tt> was executed.
3040</p>
3041
3042</div>
3043
3044<!-- _______________________________________________________________________ -->
3045<div class="doc_subsubsection">
3046 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
3047</div>
3048
3049<div class="doc_text">
3050
3051<h5>Syntax:</h5>
3052<pre>
3053 declare void %llvm.stackrestore(sbyte* %ptr)
3054</pre>
3055
3056<h5>Overview:</h5>
3057
3058<p>
3059The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
3060the function stack to the state it was in when the corresponding <a
3061href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
3062useful for implementing language features like scoped automatic variable sized
3063arrays in C99.
3064</p>
3065
3066<h5>Semantics:</h5>
3067
3068<p>
3069See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
3070</p>
3071
3072</div>
3073
3074
3075<!-- _______________________________________________________________________ -->
3076<div class="doc_subsubsection">
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003077 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
3078</div>
3079
3080<div class="doc_text">
3081
3082<h5>Syntax:</h5>
3083<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003084 declare void %llvm.prefetch(sbyte * &lt;address&gt;,
3085 uint &lt;rw&gt;, uint &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003086</pre>
3087
3088<h5>Overview:</h5>
3089
3090
3091<p>
3092The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +00003093a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
3094no
3095effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +00003096characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003097</p>
3098
3099<h5>Arguments:</h5>
3100
3101<p>
3102<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
3103determining if the fetch should be for a read (0) or write (1), and
3104<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +00003105locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003106<tt>locality</tt> arguments must be constant integers.
3107</p>
3108
3109<h5>Semantics:</h5>
3110
3111<p>
3112This intrinsic does not modify the behavior of the program. In particular,
3113prefetches cannot trap and do not produce a value. On targets that support this
3114intrinsic, the prefetch can provide hints to the processor cache for better
3115performance.
3116</p>
3117
3118</div>
3119
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003120<!-- _______________________________________________________________________ -->
3121<div class="doc_subsubsection">
3122 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
3123</div>
3124
3125<div class="doc_text">
3126
3127<h5>Syntax:</h5>
3128<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003129 declare void %llvm.pcmarker( uint &lt;id&gt; )
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003130</pre>
3131
3132<h5>Overview:</h5>
3133
3134
3135<p>
John Criswellfc6b8952005-05-16 16:17:45 +00003136The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
3137(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003138code to simulators and other tools. The method is target specific, but it is
3139expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00003140The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +00003141after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb3e7afd2006-03-24 07:16:10 +00003142optimizations. The intended use is to be inserted after optimizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +00003143correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003144</p>
3145
3146<h5>Arguments:</h5>
3147
3148<p>
3149<tt>id</tt> is a numerical id identifying the marker.
3150</p>
3151
3152<h5>Semantics:</h5>
3153
3154<p>
3155This intrinsic does not modify the behavior of the program. Backends that do not
3156support this intrinisic may ignore it.
3157</p>
3158
3159</div>
3160
Andrew Lenharth51b8d542005-11-11 16:47:30 +00003161<!-- _______________________________________________________________________ -->
3162<div class="doc_subsubsection">
3163 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
3164</div>
3165
3166<div class="doc_text">
3167
3168<h5>Syntax:</h5>
3169<pre>
3170 declare ulong %llvm.readcyclecounter( )
3171</pre>
3172
3173<h5>Overview:</h5>
3174
3175
3176<p>
3177The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
3178counter register (or similar low latency, high accuracy clocks) on those targets
3179that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
3180As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
3181should only be used for small timings.
3182</p>
3183
3184<h5>Semantics:</h5>
3185
3186<p>
3187When directly supported, reading the cycle counter should not modify any memory.
3188Implementations are allowed to either return a application specific value or a
3189system wide value. On backends without support, this is lowered to a constant 0.
3190</p>
3191
3192</div>
3193
Chris Lattner10610642004-02-14 04:08:35 +00003194<!-- ======================================================================= -->
3195<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +00003196 <a name="int_libc">Standard C Library Intrinsics</a>
3197</div>
3198
3199<div class="doc_text">
3200<p>
Chris Lattner10610642004-02-14 04:08:35 +00003201LLVM provides intrinsics for a few important standard C library functions.
3202These intrinsics allow source-language front-ends to pass information about the
3203alignment of the pointer arguments to the code generator, providing opportunity
3204for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003205</p>
3206
3207</div>
3208
3209<!-- _______________________________________________________________________ -->
3210<div class="doc_subsubsection">
3211 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
3212</div>
3213
3214<div class="doc_text">
3215
3216<h5>Syntax:</h5>
3217<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00003218 declare void %llvm.memcpy.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3219 uint &lt;len&gt;, uint &lt;align&gt;)
3220 declare void %llvm.memcpy.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3221 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003222</pre>
3223
3224<h5>Overview:</h5>
3225
3226<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003227The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00003228location to the destination location.
3229</p>
3230
3231<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003232Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
3233intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003234</p>
3235
3236<h5>Arguments:</h5>
3237
3238<p>
3239The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00003240the source. The third argument is an integer argument
Chris Lattner33aec9e2004-02-12 17:01:32 +00003241specifying the number of bytes to copy, and the fourth argument is the alignment
3242of the source and destination locations.
3243</p>
3244
Chris Lattner3301ced2004-02-12 21:18:15 +00003245<p>
3246If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00003247the caller guarantees that both the source and destination pointers are aligned
3248to that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00003249</p>
3250
Chris Lattner33aec9e2004-02-12 17:01:32 +00003251<h5>Semantics:</h5>
3252
3253<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003254The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00003255location to the destination location, which are not allowed to overlap. It
3256copies "len" bytes of memory over. If the argument is known to be aligned to
3257some boundary, this can be specified as the fourth argument, otherwise it should
3258be set to 0 or 1.
3259</p>
3260</div>
3261
3262
Chris Lattner0eb51b42004-02-12 18:10:10 +00003263<!-- _______________________________________________________________________ -->
3264<div class="doc_subsubsection">
3265 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
3266</div>
3267
3268<div class="doc_text">
3269
3270<h5>Syntax:</h5>
3271<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00003272 declare void %llvm.memmove.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3273 uint &lt;len&gt;, uint &lt;align&gt;)
3274 declare void %llvm.memmove.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3275 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00003276</pre>
3277
3278<h5>Overview:</h5>
3279
3280<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003281The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
3282location to the destination location. It is similar to the
3283'<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattner0eb51b42004-02-12 18:10:10 +00003284</p>
3285
3286<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003287Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
3288intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner0eb51b42004-02-12 18:10:10 +00003289</p>
3290
3291<h5>Arguments:</h5>
3292
3293<p>
3294The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00003295the source. The third argument is an integer argument
Chris Lattner0eb51b42004-02-12 18:10:10 +00003296specifying the number of bytes to copy, and the fourth argument is the alignment
3297of the source and destination locations.
3298</p>
3299
Chris Lattner3301ced2004-02-12 21:18:15 +00003300<p>
3301If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00003302the caller guarantees that the source and destination pointers are aligned to
3303that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00003304</p>
3305
Chris Lattner0eb51b42004-02-12 18:10:10 +00003306<h5>Semantics:</h5>
3307
3308<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003309The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner0eb51b42004-02-12 18:10:10 +00003310location to the destination location, which may overlap. It
3311copies "len" bytes of memory over. If the argument is known to be aligned to
3312some boundary, this can be specified as the fourth argument, otherwise it should
3313be set to 0 or 1.
3314</p>
3315</div>
3316
Chris Lattner8ff75902004-01-06 05:31:32 +00003317
Chris Lattner10610642004-02-14 04:08:35 +00003318<!-- _______________________________________________________________________ -->
3319<div class="doc_subsubsection">
Chris Lattner5b310c32006-03-03 00:07:20 +00003320 <a name="i_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner10610642004-02-14 04:08:35 +00003321</div>
3322
3323<div class="doc_text">
3324
3325<h5>Syntax:</h5>
3326<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00003327 declare void %llvm.memset.i32(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
3328 uint &lt;len&gt;, uint &lt;align&gt;)
3329 declare void %llvm.memset.i64(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
3330 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003331</pre>
3332
3333<h5>Overview:</h5>
3334
3335<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003336The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner10610642004-02-14 04:08:35 +00003337byte value.
3338</p>
3339
3340<p>
3341Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
3342does not return a value, and takes an extra alignment argument.
3343</p>
3344
3345<h5>Arguments:</h5>
3346
3347<p>
3348The first argument is a pointer to the destination to fill, the second is the
Chris Lattner5b310c32006-03-03 00:07:20 +00003349byte value to fill it with, the third argument is an integer
Chris Lattner10610642004-02-14 04:08:35 +00003350argument specifying the number of bytes to fill, and the fourth argument is the
3351known alignment of destination location.
3352</p>
3353
3354<p>
3355If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00003356the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner10610642004-02-14 04:08:35 +00003357</p>
3358
3359<h5>Semantics:</h5>
3360
3361<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003362The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
3363the
Chris Lattner10610642004-02-14 04:08:35 +00003364destination location. If the argument is known to be aligned to some boundary,
3365this can be specified as the fourth argument, otherwise it should be set to 0 or
33661.
3367</p>
3368</div>
3369
3370
Chris Lattner32006282004-06-11 02:28:03 +00003371<!-- _______________________________________________________________________ -->
3372<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +00003373 <a name="i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a>
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003374</div>
3375
3376<div class="doc_text">
3377
3378<h5>Syntax:</h5>
3379<pre>
Reid Spencer0b118202006-01-16 21:12:35 +00003380 declare bool %llvm.isunordered.f32(float Val1, float Val2)
3381 declare bool %llvm.isunordered.f64(double Val1, double Val2)
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003382</pre>
3383
3384<h5>Overview:</h5>
3385
3386<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003387The '<tt>llvm.isunordered</tt>' intrinsics return true if either or both of the
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00003388specified floating point values is a NAN.
3389</p>
3390
3391<h5>Arguments:</h5>
3392
3393<p>
3394The arguments are floating point numbers of the same type.
3395</p>
3396
3397<h5>Semantics:</h5>
3398
3399<p>
3400If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
3401false.
3402</p>
3403</div>
3404
3405
Chris Lattnera4d74142005-07-21 01:29:16 +00003406<!-- _______________________________________________________________________ -->
3407<div class="doc_subsubsection">
Chris Lattnerec6cb612006-01-16 22:38:59 +00003408 <a name="i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +00003409</div>
3410
3411<div class="doc_text">
3412
3413<h5>Syntax:</h5>
3414<pre>
Reid Spencer0b118202006-01-16 21:12:35 +00003415 declare double %llvm.sqrt.f32(float Val)
3416 declare double %llvm.sqrt.f64(double Val)
Chris Lattnera4d74142005-07-21 01:29:16 +00003417</pre>
3418
3419<h5>Overview:</h5>
3420
3421<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003422The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Chris Lattnera4d74142005-07-21 01:29:16 +00003423returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
3424<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
3425negative numbers (which allows for better optimization).
3426</p>
3427
3428<h5>Arguments:</h5>
3429
3430<p>
3431The argument and return value are floating point numbers of the same type.
3432</p>
3433
3434<h5>Semantics:</h5>
3435
3436<p>
3437This function returns the sqrt of the specified operand if it is a positive
3438floating point number.
3439</p>
3440</div>
3441
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003442<!-- ======================================================================= -->
3443<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +00003444 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003445</div>
3446
3447<div class="doc_text">
3448<p>
Nate Begeman7e36c472006-01-13 23:26:38 +00003449LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003450These allow efficient code generation for some algorithms.
3451</p>
3452
3453</div>
3454
3455<!-- _______________________________________________________________________ -->
3456<div class="doc_subsubsection">
Nate Begeman7e36c472006-01-13 23:26:38 +00003457 <a name="i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
3458</div>
3459
3460<div class="doc_text">
3461
3462<h5>Syntax:</h5>
3463<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003464 declare ushort %llvm.bswap.i16(ushort &lt;id&gt;)
3465 declare uint %llvm.bswap.i32(uint &lt;id&gt;)
3466 declare ulong %llvm.bswap.i64(ulong &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +00003467</pre>
3468
3469<h5>Overview:</h5>
3470
3471<p>
3472The '<tt>llvm.bwsap</tt>' family of intrinsics is used to byteswap a 16, 32 or
347364 bit quantity. These are useful for performing operations on data that is not
3474in the target's native byte order.
3475</p>
3476
3477<h5>Semantics:</h5>
3478
3479<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003480The <tt>llvm.bswap.16</tt> intrinsic returns a ushort value that has the high and low
3481byte of the input ushort swapped. Similarly, the <tt>llvm.bswap.i32</tt> intrinsic
Nate Begeman7e36c472006-01-13 23:26:38 +00003482returns a uint value that has the four bytes of the input uint swapped, so that
3483if the input bytes are numbered 0, 1, 2, 3 then the returned uint will have its
Chris Lattnerec6cb612006-01-16 22:38:59 +00003484bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i64</tt> intrinsic extends this concept
Nate Begeman7e36c472006-01-13 23:26:38 +00003485to 64 bits.
3486</p>
3487
3488</div>
3489
3490<!-- _______________________________________________________________________ -->
3491<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +00003492 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003493</div>
3494
3495<div class="doc_text">
3496
3497<h5>Syntax:</h5>
3498<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003499 declare ubyte %llvm.ctpop.i8 (ubyte &lt;src&gt;)
3500 declare ushort %llvm.ctpop.i16(ushort &lt;src&gt;)
3501 declare uint %llvm.ctpop.i32(uint &lt;src&gt;)
3502 declare ulong %llvm.ctpop.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003503</pre>
3504
3505<h5>Overview:</h5>
3506
3507<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003508The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
3509value.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003510</p>
3511
3512<h5>Arguments:</h5>
3513
3514<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003515The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00003516unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003517</p>
3518
3519<h5>Semantics:</h5>
3520
3521<p>
3522The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
3523</p>
3524</div>
3525
3526<!-- _______________________________________________________________________ -->
3527<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00003528 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003529</div>
3530
3531<div class="doc_text">
3532
3533<h5>Syntax:</h5>
3534<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003535 declare ubyte %llvm.ctlz.i8 (ubyte &lt;src&gt;)
3536 declare ushort %llvm.ctlz.i16(ushort &lt;src&gt;)
3537 declare uint %llvm.ctlz.i32(uint &lt;src&gt;)
3538 declare ulong %llvm.ctlz.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003539</pre>
3540
3541<h5>Overview:</h5>
3542
3543<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003544The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
3545leading zeros in a variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003546</p>
3547
3548<h5>Arguments:</h5>
3549
3550<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00003551The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00003552unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003553</p>
3554
3555<h5>Semantics:</h5>
3556
3557<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00003558The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
3559in a variable. If the src == 0 then the result is the size in bits of the type
3560of src. For example, <tt>llvm.cttz(int 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00003561</p>
3562</div>
Chris Lattner32006282004-06-11 02:28:03 +00003563
3564
Chris Lattnereff29ab2005-05-15 19:39:26 +00003565
3566<!-- _______________________________________________________________________ -->
3567<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00003568 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +00003569</div>
3570
3571<div class="doc_text">
3572
3573<h5>Syntax:</h5>
3574<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00003575 declare ubyte %llvm.cttz.i8 (ubyte &lt;src&gt;)
3576 declare ushort %llvm.cttz.i16(ushort &lt;src&gt;)
3577 declare uint %llvm.cttz.i32(uint &lt;src&gt;)
3578 declare ulong %llvm.cttz.i64(ulong &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +00003579</pre>
3580
3581<h5>Overview:</h5>
3582
3583<p>
Reid Spencer0b118202006-01-16 21:12:35 +00003584The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
3585trailing zeros.
Chris Lattnereff29ab2005-05-15 19:39:26 +00003586</p>
3587
3588<h5>Arguments:</h5>
3589
3590<p>
3591The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00003592unsigned integer type. The return type must match the argument type.
Chris Lattnereff29ab2005-05-15 19:39:26 +00003593</p>
3594
3595<h5>Semantics:</h5>
3596
3597<p>
3598The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
3599in a variable. If the src == 0 then the result is the size in bits of the type
3600of src. For example, <tt>llvm.cttz(2) = 1</tt>.
3601</p>
3602</div>
3603
Chris Lattner8ff75902004-01-06 05:31:32 +00003604<!-- ======================================================================= -->
3605<div class="doc_subsection">
3606 <a name="int_debugger">Debugger Intrinsics</a>
3607</div>
3608
3609<div class="doc_text">
3610<p>
3611The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
3612are described in the <a
3613href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
3614Debugging</a> document.
3615</p>
3616</div>
3617
3618
Chris Lattner00950542001-06-06 20:29:01 +00003619<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00003620<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00003621<address>
3622 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3623 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3624 <a href="http://validator.w3.org/check/referer"><img
3625 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3626
3627 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +00003628 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00003629 Last modified: $Date$
3630</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003631</body>
3632</html>