blob: 7f0353102faf7e493e768d171600e76471ec9c1f [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>
Reid Spencer1628cec2006-10-26 06:15:43 +000080 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
81 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
82 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +000083 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
84 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
85 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000086 </ol>
87 </li>
Chris Lattner00950542001-06-06 20:29:01 +000088 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
89 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +000090 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000091 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000092 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
93 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
Reid Spencer3822ff52006-11-08 06:47:33 +000094 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
95 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000096 </ol>
97 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +000098 <li><a href="#vectorops">Vector Operations</a>
99 <ol>
100 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
101 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
102 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000103 </ol>
104 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000105 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000106 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000107 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
108 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
109 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000110 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
111 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
112 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000113 </ol>
114 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000115 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000116 <ol>
117 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
118 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
119 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
120 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
121 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000122 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
123 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
124 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
125 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000126 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
127 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000128 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000129 </ol>
Chris Lattner00950542001-06-06 20:29:01 +0000130 <li><a href="#otherops">Other Operations</a>
131 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000132 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
133 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000134 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000135 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000136 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000137 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000138 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000139 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000140 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000141 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000142 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000143 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000144 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
145 <ol>
146 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
147 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
148 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
149 </ol>
150 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000151 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
152 <ol>
153 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
154 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
155 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
156 </ol>
157 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000158 <li><a href="#int_codegen">Code Generator Intrinsics</a>
159 <ol>
160 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
161 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
Chris Lattner57e1f392006-01-13 02:03:13 +0000162 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
163 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000164 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000165 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000166 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000167 </ol>
168 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000169 <li><a href="#int_libc">Standard C Library Intrinsics</a>
170 <ol>
Chris Lattner5b310c32006-03-03 00:07:20 +0000171 <li><a href="#i_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
172 <li><a href="#i_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
173 <li><a href="#i_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000174 <li><a href="#i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a></li>
175 <li><a href="#i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000176 <li><a href="#i_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000177 </ol>
178 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000179 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000180 <ol>
Nate Begeman7e36c472006-01-13 23:26:38 +0000181 <li><a href="#i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000182 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
183 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
184 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000185 </ol>
186 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000187 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000188 </ol>
189 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000190</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000191
192<div class="doc_author">
193 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
194 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000195</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000196
Chris Lattner00950542001-06-06 20:29:01 +0000197<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000198<div class="doc_section"> <a name="abstract">Abstract </a></div>
199<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000200
Misha Brukman9d0919f2003-11-08 01:05:38 +0000201<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000202<p>This document is a reference manual for the LLVM assembly language.
203LLVM is an SSA based representation that provides type safety,
204low-level operations, flexibility, and the capability of representing
205'all' high-level languages cleanly. It is the common code
206representation used throughout all phases of the LLVM compilation
207strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000208</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000209
Chris Lattner00950542001-06-06 20:29:01 +0000210<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000211<div class="doc_section"> <a name="introduction">Introduction</a> </div>
212<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000213
Misha Brukman9d0919f2003-11-08 01:05:38 +0000214<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000215
Chris Lattner261efe92003-11-25 01:02:51 +0000216<p>The LLVM code representation is designed to be used in three
217different forms: as an in-memory compiler IR, as an on-disk bytecode
218representation (suitable for fast loading by a Just-In-Time compiler),
219and as a human readable assembly language representation. This allows
220LLVM to provide a powerful intermediate representation for efficient
221compiler transformations and analysis, while providing a natural means
222to debug and visualize the transformations. The three different forms
223of LLVM are all equivalent. This document describes the human readable
224representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000225
John Criswellc1f786c2005-05-13 22:25:59 +0000226<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000227while being expressive, typed, and extensible at the same time. It
228aims to be a "universal IR" of sorts, by being at a low enough level
229that high-level ideas may be cleanly mapped to it (similar to how
230microprocessors are "universal IR's", allowing many source languages to
231be mapped to them). By providing type information, LLVM can be used as
232the target of optimizations: for example, through pointer analysis, it
233can be proven that a C automatic variable is never accessed outside of
234the current function... allowing it to be promoted to a simple SSA
235value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000236
Misha Brukman9d0919f2003-11-08 01:05:38 +0000237</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000238
Chris Lattner00950542001-06-06 20:29:01 +0000239<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000240<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000241
Misha Brukman9d0919f2003-11-08 01:05:38 +0000242<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000243
Chris Lattner261efe92003-11-25 01:02:51 +0000244<p>It is important to note that this document describes 'well formed'
245LLVM assembly language. There is a difference between what the parser
246accepts and what is considered 'well formed'. For example, the
247following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000248
249<pre>
250 %x = <a href="#i_add">add</a> int 1, %x
251</pre>
252
Chris Lattner261efe92003-11-25 01:02:51 +0000253<p>...because the definition of <tt>%x</tt> does not dominate all of
254its uses. The LLVM infrastructure provides a verification pass that may
255be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000256automatically run by the parser after parsing input assembly and by
Chris Lattner261efe92003-11-25 01:02:51 +0000257the optimizer before it outputs bytecode. The violations pointed out
258by the verifier pass indicate bugs in transformation passes or input to
259the parser.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000260
Chris Lattner261efe92003-11-25 01:02:51 +0000261<!-- Describe the typesetting conventions here. --> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000262
Chris Lattner00950542001-06-06 20:29:01 +0000263<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000264<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000265<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000266
Misha Brukman9d0919f2003-11-08 01:05:38 +0000267<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000268
Chris Lattner261efe92003-11-25 01:02:51 +0000269<p>LLVM uses three different forms of identifiers, for different
270purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000271
Chris Lattner00950542001-06-06 20:29:01 +0000272<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000273 <li>Named values are represented as a string of characters with a '%' prefix.
274 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
275 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
276 Identifiers which require other characters in their names can be surrounded
277 with quotes. In this way, anything except a <tt>"</tt> character can be used
278 in a name.</li>
279
280 <li>Unnamed values are represented as an unsigned numeric value with a '%'
281 prefix. For example, %12, %2, %44.</li>
282
Reid Spencercc16dc32004-12-09 18:02:53 +0000283 <li>Constants, which are described in a <a href="#constants">section about
284 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000285</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000286
287<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
288don't need to worry about name clashes with reserved words, and the set of
289reserved words may be expanded in the future without penalty. Additionally,
290unnamed identifiers allow a compiler to quickly come up with a temporary
291variable without having to avoid symbol table conflicts.</p>
292
Chris Lattner261efe92003-11-25 01:02:51 +0000293<p>Reserved words in LLVM are very similar to reserved words in other
Reid Spencer5c0ef472006-11-11 23:08:07 +0000294languages. There are keywords for different opcodes
295('<tt><a href="#i_add">add</a></tt>',
296 '<tt><a href="#i_bitcast">bitcast</a></tt>',
297 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
Chris Lattnere5d947b2004-12-09 16:36:40 +0000298href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
299and others. These reserved words cannot conflict with variable names, because
300none of them start with a '%' character.</p>
301
302<p>Here is an example of LLVM code to multiply the integer variable
303'<tt>%X</tt>' by 8:</p>
304
Misha Brukman9d0919f2003-11-08 01:05:38 +0000305<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000306
307<pre>
308 %result = <a href="#i_mul">mul</a> uint %X, 8
309</pre>
310
Misha Brukman9d0919f2003-11-08 01:05:38 +0000311<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000312
313<pre>
314 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
315</pre>
316
Misha Brukman9d0919f2003-11-08 01:05:38 +0000317<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000318
319<pre>
320 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
321 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
322 %result = <a href="#i_add">add</a> uint %1, %1
323</pre>
324
Chris Lattner261efe92003-11-25 01:02:51 +0000325<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
326important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000327
Chris Lattner00950542001-06-06 20:29:01 +0000328<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000329
330 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
331 line.</li>
332
333 <li>Unnamed temporaries are created when the result of a computation is not
334 assigned to a named value.</li>
335
Misha Brukman9d0919f2003-11-08 01:05:38 +0000336 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000337
Misha Brukman9d0919f2003-11-08 01:05:38 +0000338</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000339
John Criswelle4c57cc2005-05-12 16:52:32 +0000340<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000341demonstrating instructions, we will follow an instruction with a comment that
342defines the type and name of value produced. Comments are shown in italic
343text.</p>
344
Misha Brukman9d0919f2003-11-08 01:05:38 +0000345</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000346
347<!-- *********************************************************************** -->
348<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
349<!-- *********************************************************************** -->
350
351<!-- ======================================================================= -->
352<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
353</div>
354
355<div class="doc_text">
356
357<p>LLVM programs are composed of "Module"s, each of which is a
358translation unit of the input programs. Each module consists of
359functions, global variables, and symbol table entries. Modules may be
360combined together with the LLVM linker, which merges function (and
361global variable) definitions, resolves forward declarations, and merges
362symbol table entries. Here is an example of the "hello world" module:</p>
363
364<pre><i>; Declare the string constant as a global constant...</i>
365<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
366 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
367
368<i>; External declaration of the puts function</i>
369<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
370
Chris Lattner81c01f02006-06-13 03:05:47 +0000371<i>; Global variable / Function body section separator</i>
372implementation
373
Chris Lattnerfa730212004-12-09 16:11:40 +0000374<i>; Definition of main function</i>
375int %main() { <i>; int()* </i>
376 <i>; Convert [13x sbyte]* to sbyte *...</i>
377 %cast210 = <a
378 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
379
380 <i>; Call puts function to write out the string to stdout...</i>
381 <a
382 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
383 <a
384 href="#i_ret">ret</a> int 0<br>}<br></pre>
385
386<p>This example is made up of a <a href="#globalvars">global variable</a>
387named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
388function, and a <a href="#functionstructure">function definition</a>
389for "<tt>main</tt>".</p>
390
Chris Lattnere5d947b2004-12-09 16:36:40 +0000391<p>In general, a module is made up of a list of global values,
392where both functions and global variables are global values. Global values are
393represented by a pointer to a memory location (in this case, a pointer to an
394array of char, and a pointer to a function), and have one of the following <a
395href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000396
Chris Lattner81c01f02006-06-13 03:05:47 +0000397<p>Due to a limitation in the current LLVM assembly parser (it is limited by
398one-token lookahead), modules are split into two pieces by the "implementation"
399keyword. Global variable prototypes and definitions must occur before the
400keyword, and function definitions must occur after it. Function prototypes may
401occur either before or after it. In the future, the implementation keyword may
402become a noop, if the parser gets smarter.</p>
403
Chris Lattnere5d947b2004-12-09 16:36:40 +0000404</div>
405
406<!-- ======================================================================= -->
407<div class="doc_subsection">
408 <a name="linkage">Linkage Types</a>
409</div>
410
411<div class="doc_text">
412
413<p>
414All Global Variables and Functions have one of the following types of linkage:
415</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000416
417<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000418
Chris Lattnerfa730212004-12-09 16:11:40 +0000419 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000420
421 <dd>Global values with internal linkage are only directly accessible by
422 objects in the current module. In particular, linking code into a module with
423 an internal global value may cause the internal to be renamed as necessary to
424 avoid collisions. Because the symbol is internal to the module, all
425 references can be updated. This corresponds to the notion of the
426 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattnerfa730212004-12-09 16:11:40 +0000427 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000428
Chris Lattnerfa730212004-12-09 16:11:40 +0000429 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000430
431 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
432 the twist that linking together two modules defining the same
433 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
434 is typically used to implement inline functions. Unreferenced
435 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000436 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000437
Chris Lattnerfa730212004-12-09 16:11:40 +0000438 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000439
440 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
441 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
442 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000443 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000444
Chris Lattnerfa730212004-12-09 16:11:40 +0000445 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000446
447 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
448 pointer to array type. When two global variables with appending linkage are
449 linked together, the two global arrays are appended together. This is the
450 LLVM, typesafe, equivalent of having the system linker append together
451 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000452 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000453
Chris Lattnerfa730212004-12-09 16:11:40 +0000454 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000455
456 <dd>If none of the above identifiers are used, the global is externally
457 visible, meaning that it participates in linkage and can be used to resolve
458 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000459 </dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000460
461 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
462
463 <dd>"<tt>extern_weak</tt>" TBD
464 </dd>
465
466 <p>
467 The next two types of linkage are targeted for Microsoft Windows platform
468 only. They are designed to support importing (exporting) symbols from (to)
469 DLLs.
470 </p>
471
472 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
473
474 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
475 or variable via a global pointer to a pointer that is set up by the DLL
476 exporting the symbol. On Microsoft Windows targets, the pointer name is
477 formed by combining <code>_imp__</code> and the function or variable name.
478 </dd>
479
480 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
481
482 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
483 pointer to a pointer in a DLL, so that it can be referenced with the
484 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
485 name is formed by combining <code>_imp__</code> and the function or variable
486 name.
487 </dd>
488
Chris Lattnerfa730212004-12-09 16:11:40 +0000489</dl>
490
Chris Lattnerfa730212004-12-09 16:11:40 +0000491<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
492variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
493variable and was linked with this one, one of the two would be renamed,
494preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
495external (i.e., lacking any linkage declarations), they are accessible
496outside of the current module. It is illegal for a function <i>declaration</i>
497to have any linkage type other than "externally visible".</a></p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000498
Chris Lattnerfa730212004-12-09 16:11:40 +0000499</div>
500
501<!-- ======================================================================= -->
502<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000503 <a name="callingconv">Calling Conventions</a>
504</div>
505
506<div class="doc_text">
507
508<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
509and <a href="#i_invoke">invokes</a> can all have an optional calling convention
510specified for the call. The calling convention of any pair of dynamic
511caller/callee must match, or the behavior of the program is undefined. The
512following calling conventions are supported by LLVM, and more may be added in
513the future:</p>
514
515<dl>
516 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
517
518 <dd>This calling convention (the default if no other calling convention is
519 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000520 supports varargs function calls and tolerates some mismatch in the declared
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000521 prototype and implemented declaration of the function (as does normal C).
522 </dd>
523
Chris Lattner5710ce92006-05-19 21:15:36 +0000524 <dt><b>"<tt>csretcc</tt>" - The C struct return calling convention</b>:</dt>
525
526 <dd>This calling convention matches the target C calling conventions, except
527 that functions with this convention are required to take a pointer as their
528 first argument, and the return type of the function must be void. This is
529 used for C functions that return aggregates by-value. In this case, the
530 function has been transformed to take a pointer to the struct as the first
531 argument to the function. For targets where the ABI specifies specific
532 behavior for structure-return calls, the calling convention can be used to
533 distinguish between struct return functions and other functions that take a
534 pointer to a struct as the first argument.
535 </dd>
536
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000537 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
538
539 <dd>This calling convention attempts to make calls as fast as possible
540 (e.g. by passing things in registers). This calling convention allows the
541 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000542 without having to conform to an externally specified ABI. Implementations of
543 this convention should allow arbitrary tail call optimization to be supported.
544 This calling convention does not support varargs and requires the prototype of
545 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000546 </dd>
547
548 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
549
550 <dd>This calling convention attempts to make code in the caller as efficient
551 as possible under the assumption that the call is not commonly executed. As
552 such, these calls often preserve all registers so that the call does not break
553 any live ranges in the caller side. This calling convention does not support
554 varargs and requires the prototype of all callees to exactly match the
555 prototype of the function definition.
556 </dd>
557
Chris Lattnercfe6b372005-05-07 01:46:40 +0000558 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000559
560 <dd>Any calling convention may be specified by number, allowing
561 target-specific calling conventions to be used. Target specific calling
562 conventions start at 64.
563 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000564</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000565
566<p>More calling conventions can be added/defined on an as-needed basis, to
567support pascal conventions or any other well-known target-independent
568convention.</p>
569
570</div>
571
572<!-- ======================================================================= -->
573<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000574 <a name="globalvars">Global Variables</a>
575</div>
576
577<div class="doc_text">
578
Chris Lattner3689a342005-02-12 19:30:21 +0000579<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000580instead of run-time. Global variables may optionally be initialized, may have
581an explicit section to be placed in, and may
Chris Lattner2cbdc452005-11-06 08:02:57 +0000582have an optional explicit alignment specified. A
John Criswell0ec250c2005-10-24 16:17:18 +0000583variable may be defined as a global "constant," which indicates that the
Chris Lattner3689a342005-02-12 19:30:21 +0000584contents of the variable will <b>never</b> be modified (enabling better
585optimization, allowing the global data to be placed in the read-only section of
586an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000587cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000588
589<p>
590LLVM explicitly allows <em>declarations</em> of global variables to be marked
591constant, even if the final definition of the global is not. This capability
592can be used to enable slightly better optimization of the program, but requires
593the language definition to guarantee that optimizations based on the
594'constantness' are valid for the translation units that do not include the
595definition.
596</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000597
598<p>As SSA values, global variables define pointer values that are in
599scope (i.e. they dominate) all basic blocks in the program. Global
600variables always define a pointer to their "content" type because they
601describe a region of memory, and all memory objects in LLVM are
602accessed through pointers.</p>
603
Chris Lattner88f6c462005-11-12 00:45:07 +0000604<p>LLVM allows an explicit section to be specified for globals. If the target
605supports it, it will emit globals to the section specified.</p>
606
Chris Lattner2cbdc452005-11-06 08:02:57 +0000607<p>An explicit alignment may be specified for a global. If not present, or if
608the alignment is set to zero, the alignment of the global is set by the target
609to whatever it feels convenient. If an explicit alignment is specified, the
610global is forced to have at least that much alignment. All alignments must be
611a power of 2.</p>
612
Chris Lattnerfa730212004-12-09 16:11:40 +0000613</div>
614
615
616<!-- ======================================================================= -->
617<div class="doc_subsection">
618 <a name="functionstructure">Functions</a>
619</div>
620
621<div class="doc_text">
622
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000623<p>LLVM function definitions consist of an optional <a href="#linkage">linkage
624type</a>, an optional <a href="#callingconv">calling convention</a>, a return
Chris Lattner88f6c462005-11-12 00:45:07 +0000625type, a function name, a (possibly empty) argument list, an optional section,
626an optional alignment, an opening curly brace,
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000627a list of basic blocks, and a closing curly brace. LLVM function declarations
628are defined with the "<tt>declare</tt>" keyword, an optional <a
Chris Lattner2cbdc452005-11-06 08:02:57 +0000629href="#callingconv">calling convention</a>, a return type, a function name,
630a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000631
632<p>A function definition contains a list of basic blocks, forming the CFG for
633the function. Each basic block may optionally start with a label (giving the
634basic block a symbol table entry), contains a list of instructions, and ends
635with a <a href="#terminators">terminator</a> instruction (such as a branch or
636function return).</p>
637
John Criswelle4c57cc2005-05-12 16:52:32 +0000638<p>The first basic block in a program is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000639executed on entrance to the function, and it is not allowed to have predecessor
640basic blocks (i.e. there can not be any branches to the entry block of a
641function). Because the block can have no predecessors, it also cannot have any
642<a href="#i_phi">PHI nodes</a>.</p>
643
644<p>LLVM functions are identified by their name and type signature. Hence, two
645functions with the same name but different parameter lists or return values are
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000646considered different functions, and LLVM will resolve references to each
Chris Lattnerfa730212004-12-09 16:11:40 +0000647appropriately.</p>
648
Chris Lattner88f6c462005-11-12 00:45:07 +0000649<p>LLVM allows an explicit section to be specified for functions. If the target
650supports it, it will emit functions to the section specified.</p>
651
Chris Lattner2cbdc452005-11-06 08:02:57 +0000652<p>An explicit alignment may be specified for a function. If not present, or if
653the alignment is set to zero, the alignment of the function is set by the target
654to whatever it feels convenient. If an explicit alignment is specified, the
655function is forced to have at least that much alignment. All alignments must be
656a power of 2.</p>
657
Chris Lattnerfa730212004-12-09 16:11:40 +0000658</div>
659
Chris Lattner4e9aba72006-01-23 23:23:47 +0000660<!-- ======================================================================= -->
661<div class="doc_subsection">
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000662 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000663</div>
664
665<div class="doc_text">
666<p>
667Modules may contain "module-level inline asm" blocks, which corresponds to the
668GCC "file scope inline asm" blocks. These blocks are internally concatenated by
669LLVM and treated as a single unit, but may be separated in the .ll file if
670desired. The syntax is very simple:
671</p>
672
673<div class="doc_code"><pre>
Chris Lattner52599e12006-01-24 00:37:20 +0000674 module asm "inline asm code goes here"
675 module asm "more can go here"
Chris Lattner4e9aba72006-01-23 23:23:47 +0000676</pre></div>
677
678<p>The strings can contain any character by escaping non-printable characters.
679 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
680 for the number.
681</p>
682
683<p>
684 The inline asm code is simply printed to the machine code .s file when
685 assembly code is generated.
686</p>
687</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000688
689
Chris Lattner00950542001-06-06 20:29:01 +0000690<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000691<div class="doc_section"> <a name="typesystem">Type System</a> </div>
692<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000693
Misha Brukman9d0919f2003-11-08 01:05:38 +0000694<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000695
Misha Brukman9d0919f2003-11-08 01:05:38 +0000696<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000697intermediate representation. Being typed enables a number of
698optimizations to be performed on the IR directly, without having to do
699extra analyses on the side before the transformation. A strong type
700system makes it easier to read the generated code and enables novel
701analyses and transformations that are not feasible to perform on normal
702three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000703
704</div>
705
Chris Lattner00950542001-06-06 20:29:01 +0000706<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000707<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000708<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000709<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000710system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000711
Reid Spencerd3f876c2004-11-01 08:19:36 +0000712<table class="layout">
713 <tr class="layout">
714 <td class="left">
715 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000716 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000717 <tr><th>Type</th><th>Description</th></tr>
718 <tr><td><tt>void</tt></td><td>No value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000719 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
720 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
721 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
722 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
723 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000724 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000725 </tbody>
726 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000727 </td>
728 <td class="right">
729 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000730 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000731 <tr><th>Type</th><th>Description</th></tr>
732 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000733 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
734 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
735 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
736 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
737 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000738 </tbody>
739 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000740 </td>
741 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000742</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000743</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000744
Chris Lattner00950542001-06-06 20:29:01 +0000745<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000746<div class="doc_subsubsection"> <a name="t_classifications">Type
747Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000748<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000749<p>These different primitive types fall into a few useful
750classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000751
752<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000753 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000754 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000755 <tr>
756 <td><a name="t_signed">signed</a></td>
757 <td><tt>sbyte, short, int, long, float, double</tt></td>
758 </tr>
759 <tr>
760 <td><a name="t_unsigned">unsigned</a></td>
761 <td><tt>ubyte, ushort, uint, ulong</tt></td>
762 </tr>
763 <tr>
764 <td><a name="t_integer">integer</a></td>
765 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
766 </tr>
767 <tr>
768 <td><a name="t_integral">integral</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000769 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
770 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000771 </tr>
772 <tr>
773 <td><a name="t_floating">floating point</a></td>
774 <td><tt>float, double</tt></td>
775 </tr>
776 <tr>
777 <td><a name="t_firstclass">first class</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000778 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
779 float, double, <a href="#t_pointer">pointer</a>,
780 <a href="#t_packed">packed</a></tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000781 </tr>
782 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000783</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000784
Chris Lattner261efe92003-11-25 01:02:51 +0000785<p>The <a href="#t_firstclass">first class</a> types are perhaps the
786most important. Values of these types are the only ones which can be
787produced by instructions, passed as arguments, or used as operands to
788instructions. This means that all structures and arrays must be
789manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000790</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000791
Chris Lattner00950542001-06-06 20:29:01 +0000792<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000793<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000794
Misha Brukman9d0919f2003-11-08 01:05:38 +0000795<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000796
Chris Lattner261efe92003-11-25 01:02:51 +0000797<p>The real power in LLVM comes from the derived types in the system.
798This is what allows a programmer to represent arrays, functions,
799pointers, and other useful types. Note that these derived types may be
800recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000801
Misha Brukman9d0919f2003-11-08 01:05:38 +0000802</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000803
Chris Lattner00950542001-06-06 20:29:01 +0000804<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000805<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000806
Misha Brukman9d0919f2003-11-08 01:05:38 +0000807<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000808
Chris Lattner00950542001-06-06 20:29:01 +0000809<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000810
Misha Brukman9d0919f2003-11-08 01:05:38 +0000811<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000812sequentially in memory. The array type requires a size (number of
813elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000814
Chris Lattner7faa8832002-04-14 06:13:44 +0000815<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000816
817<pre>
818 [&lt;# elements&gt; x &lt;elementtype&gt;]
819</pre>
820
John Criswelle4c57cc2005-05-12 16:52:32 +0000821<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000822be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000823
Chris Lattner7faa8832002-04-14 06:13:44 +0000824<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000825<table class="layout">
826 <tr class="layout">
827 <td class="left">
828 <tt>[40 x int ]</tt><br/>
829 <tt>[41 x int ]</tt><br/>
830 <tt>[40 x uint]</tt><br/>
831 </td>
832 <td class="left">
833 Array of 40 integer values.<br/>
834 Array of 41 integer values.<br/>
835 Array of 40 unsigned integer values.<br/>
836 </td>
837 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000838</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000839<p>Here are some examples of multidimensional arrays:</p>
840<table class="layout">
841 <tr class="layout">
842 <td class="left">
843 <tt>[3 x [4 x int]]</tt><br/>
844 <tt>[12 x [10 x float]]</tt><br/>
845 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
846 </td>
847 <td class="left">
John Criswellc1f786c2005-05-13 22:25:59 +0000848 3x4 array of integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000849 12x10 array of single precision floating point values.<br/>
850 2x3x4 array of unsigned integer values.<br/>
851 </td>
852 </tr>
853</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000854
John Criswell0ec250c2005-10-24 16:17:18 +0000855<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
856length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000857LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
858As a special case, however, zero length arrays are recognized to be variable
859length. This allows implementation of 'pascal style arrays' with the LLVM
860type "{ int, [0 x float]}", for example.</p>
861
Misha Brukman9d0919f2003-11-08 01:05:38 +0000862</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000863
Chris Lattner00950542001-06-06 20:29:01 +0000864<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000865<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000866<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000867<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000868<p>The function type can be thought of as a function signature. It
869consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000870Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000871(which are structures of pointers to functions), for indirect function
872calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000873<p>
874The return type of a function type cannot be an aggregate type.
875</p>
Chris Lattner00950542001-06-06 20:29:01 +0000876<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000877<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000878<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +0000879specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000880which indicates that the function takes a variable number of arguments.
881Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000882 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000883<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000884<table class="layout">
885 <tr class="layout">
886 <td class="left">
887 <tt>int (int)</tt> <br/>
888 <tt>float (int, int *) *</tt><br/>
889 <tt>int (sbyte *, ...)</tt><br/>
890 </td>
891 <td class="left">
892 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
893 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukmanc24b7582004-08-12 20:16:08 +0000894 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
Reid Spencerd3f876c2004-11-01 08:19:36 +0000895 returning <tt>float</tt>.<br/>
896 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
897 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
898 the signature for <tt>printf</tt> in LLVM.<br/>
899 </td>
900 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000901</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000902
Misha Brukman9d0919f2003-11-08 01:05:38 +0000903</div>
Chris Lattner00950542001-06-06 20:29:01 +0000904<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000905<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000906<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000907<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000908<p>The structure type is used to represent a collection of data members
909together in memory. The packing of the field types is defined to match
910the ABI of the underlying processor. The elements of a structure may
911be any type that has a size.</p>
912<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
913and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
914field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
915instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000916<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000917<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000918<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000919<table class="layout">
920 <tr class="layout">
921 <td class="left">
922 <tt>{ int, int, int }</tt><br/>
923 <tt>{ float, int (int) * }</tt><br/>
924 </td>
925 <td class="left">
926 a triple of three <tt>int</tt> values<br/>
927 A pair, where the first element is a <tt>float</tt> and the second element
928 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
929 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
930 </td>
931 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000932</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000933</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000934
Chris Lattner00950542001-06-06 20:29:01 +0000935<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000936<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000937<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000938<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000939<p>As in many languages, the pointer type represents a pointer or
940reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000941<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000942<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000943<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000944<table class="layout">
945 <tr class="layout">
946 <td class="left">
947 <tt>[4x int]*</tt><br/>
948 <tt>int (int *) *</tt><br/>
949 </td>
950 <td class="left">
951 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
952 four <tt>int</tt> values<br/>
953 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattnera977c482005-02-19 02:22:14 +0000954 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerd3f876c2004-11-01 08:19:36 +0000955 <tt>int</tt>.<br/>
956 </td>
957 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000958</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000959</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000960
Chris Lattnera58561b2004-08-12 19:12:28 +0000961<!-- _______________________________________________________________________ -->
962<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000963<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000964
Chris Lattnera58561b2004-08-12 19:12:28 +0000965<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000966
Chris Lattnera58561b2004-08-12 19:12:28 +0000967<p>A packed type is a simple derived type that represents a vector
968of elements. Packed types are used when multiple primitive data
969are operated in parallel using a single instruction (SIMD).
970A packed type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000971elements) and an underlying primitive data type. Vectors must have a power
972of two length (1, 2, 4, 8, 16 ...). Packed types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000973considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000974
Chris Lattnera58561b2004-08-12 19:12:28 +0000975<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000976
977<pre>
978 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
979</pre>
980
John Criswellc1f786c2005-05-13 22:25:59 +0000981<p>The number of elements is a constant integer value; elementtype may
Chris Lattnera58561b2004-08-12 19:12:28 +0000982be any integral or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000983
Chris Lattnera58561b2004-08-12 19:12:28 +0000984<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000985
Reid Spencerd3f876c2004-11-01 08:19:36 +0000986<table class="layout">
987 <tr class="layout">
988 <td class="left">
989 <tt>&lt;4 x int&gt;</tt><br/>
990 <tt>&lt;8 x float&gt;</tt><br/>
991 <tt>&lt;2 x uint&gt;</tt><br/>
992 </td>
993 <td class="left">
994 Packed vector of 4 integer values.<br/>
995 Packed vector of 8 floating-point values.<br/>
996 Packed vector of 2 unsigned integer values.<br/>
997 </td>
998 </tr>
999</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001000</div>
1001
Chris Lattner69c11bb2005-04-25 17:34:15 +00001002<!-- _______________________________________________________________________ -->
1003<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1004<div class="doc_text">
1005
1006<h5>Overview:</h5>
1007
1008<p>Opaque types are used to represent unknown types in the system. This
1009corresponds (for example) to the C notion of a foward declared structure type.
1010In LLVM, opaque types can eventually be resolved to any type (not just a
1011structure type).</p>
1012
1013<h5>Syntax:</h5>
1014
1015<pre>
1016 opaque
1017</pre>
1018
1019<h5>Examples:</h5>
1020
1021<table class="layout">
1022 <tr class="layout">
1023 <td class="left">
1024 <tt>opaque</tt>
1025 </td>
1026 <td class="left">
1027 An opaque type.<br/>
1028 </td>
1029 </tr>
1030</table>
1031</div>
1032
1033
Chris Lattnerc3f59762004-12-09 17:30:23 +00001034<!-- *********************************************************************** -->
1035<div class="doc_section"> <a name="constants">Constants</a> </div>
1036<!-- *********************************************************************** -->
1037
1038<div class="doc_text">
1039
1040<p>LLVM has several different basic types of constants. This section describes
1041them all and their syntax.</p>
1042
1043</div>
1044
1045<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +00001046<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001047
1048<div class="doc_text">
1049
1050<dl>
1051 <dt><b>Boolean constants</b></dt>
1052
1053 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1054 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
1055 </dd>
1056
1057 <dt><b>Integer constants</b></dt>
1058
Reid Spencercc16dc32004-12-09 18:02:53 +00001059 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattnerc3f59762004-12-09 17:30:23 +00001060 href="#t_integer">integer</a> type. Negative numbers may be used with signed
1061 integer types.
1062 </dd>
1063
1064 <dt><b>Floating point constants</b></dt>
1065
1066 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1067 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +00001068 notation (see below). Floating point constants must have a <a
1069 href="#t_floating">floating point</a> type. </dd>
1070
1071 <dt><b>Null pointer constants</b></dt>
1072
John Criswell9e2485c2004-12-10 15:51:16 +00001073 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +00001074 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1075
1076</dl>
1077
John Criswell9e2485c2004-12-10 15:51:16 +00001078<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +00001079of floating point constants. For example, the form '<tt>double
10800x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
10814.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +00001082(and the only time that they are generated by the disassembler) is when a
1083floating point constant must be emitted but it cannot be represented as a
1084decimal floating point number. For example, NaN's, infinities, and other
1085special values are represented in their IEEE hexadecimal format so that
1086assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001087
1088</div>
1089
1090<!-- ======================================================================= -->
1091<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1092</div>
1093
1094<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001095<p>Aggregate constants arise from aggregation of simple constants
1096and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001097
1098<dl>
1099 <dt><b>Structure constants</b></dt>
1100
1101 <dd>Structure constants are represented with notation similar to structure
1102 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001103 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
1104 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
1105 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +00001106 types of elements must match those specified by the type.
1107 </dd>
1108
1109 <dt><b>Array constants</b></dt>
1110
1111 <dd>Array constants are represented with notation similar to array type
1112 definitions (a comma separated list of elements, surrounded by square brackets
John Criswell9e2485c2004-12-10 15:51:16 +00001113 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +00001114 constants must have <a href="#t_array">array type</a>, and the number and
1115 types of elements must match those specified by the type.
1116 </dd>
1117
1118 <dt><b>Packed constants</b></dt>
1119
1120 <dd>Packed constants are represented with notation similar to packed type
1121 definitions (a comma separated list of elements, surrounded by
John Criswell9e2485c2004-12-10 15:51:16 +00001122 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; int 42,
Chris Lattnerc3f59762004-12-09 17:30:23 +00001123 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
1124 href="#t_packed">packed type</a>, and the number and types of elements must
1125 match those specified by the type.
1126 </dd>
1127
1128 <dt><b>Zero initialization</b></dt>
1129
1130 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1131 value to zero of <em>any</em> type, including scalar and aggregate types.
1132 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +00001133 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +00001134 initializers.
1135 </dd>
1136</dl>
1137
1138</div>
1139
1140<!-- ======================================================================= -->
1141<div class="doc_subsection">
1142 <a name="globalconstants">Global Variable and Function Addresses</a>
1143</div>
1144
1145<div class="doc_text">
1146
1147<p>The addresses of <a href="#globalvars">global variables</a> and <a
1148href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +00001149constants. These constants are explicitly referenced when the <a
1150href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +00001151href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1152file:</p>
1153
1154<pre>
1155 %X = global int 17
1156 %Y = global int 42
1157 %Z = global [2 x int*] [ int* %X, int* %Y ]
1158</pre>
1159
1160</div>
1161
1162<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +00001163<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001164<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +00001165 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +00001166 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +00001167 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001168
Reid Spencer2dc45b82004-12-09 18:13:12 +00001169 <p>Undefined values indicate to the compiler that the program is well defined
1170 no matter what value is used, giving the compiler more freedom to optimize.
1171 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001172</div>
1173
1174<!-- ======================================================================= -->
1175<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1176</div>
1177
1178<div class="doc_text">
1179
1180<p>Constant expressions are used to allow expressions involving other constants
1181to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +00001182href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +00001183that does not have side effects (e.g. load and call are not supported). The
1184following is the syntax for constant expressions:</p>
1185
1186<dl>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001187 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1188 <dd>Truncate a constant to another type. The bit size of CST must be larger
1189 than the bit size of TYPE. Both types must be integral.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001190
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001191 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1192 <dd>Zero extend a constant to another type. The bit size of CST must be
1193 smaller or equal to the bit size of TYPE. Both types must be integral.</dd>
1194
1195 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1196 <dd>Sign extend a constant to another type. The bit size of CST must be
1197 smaller or equal to the bit size of TYPE. Both types must be integral.</dd>
1198
1199 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1200 <dd>Truncate a floating point constant to another floating point type. The
1201 size of CST must be larger than the size of TYPE. Both types must be
1202 floating point.</dd>
1203
1204 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1205 <dd>Floating point extend a constant to another type. The size of CST must be
1206 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1207
1208 <dt><b><tt>fp2uint ( CST to TYPE )</tt></b></dt>
1209 <dd>Convert a floating point constant to the corresponding unsigned integer
1210 constant. TYPE must be an integer type. CST must be floating point. If the
1211 value won't fit in the integer type, the results are undefined.</dd>
1212
Reid Spencerd4448792006-11-09 23:03:26 +00001213 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001214 <dd>Convert a floating point constant to the corresponding signed integer
1215 constant. TYPE must be an integer type. CST must be floating point. If the
1216 value won't fit in the integer type, the results are undefined.</dd>
1217
Reid Spencerd4448792006-11-09 23:03:26 +00001218 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001219 <dd>Convert an unsigned integer constant to the corresponding floating point
1220 constant. TYPE must be floating point. CST must be of integer type. If the
1221 value won't fit in the floating point type, the results are undefined.</dd>
1222
Reid Spencerd4448792006-11-09 23:03:26 +00001223 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001224 <dd>Convert a signed integer constant to the corresponding floating point
1225 constant. TYPE must be floating point. CST must be of integer type. If the
1226 value won't fit in the floating point type, the results are undefined.</dd>
1227
Reid Spencer5c0ef472006-11-11 23:08:07 +00001228 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1229 <dd>Convert a pointer typed constant to the corresponding integer constant
1230 TYPE must be an integer type. CST must be of pointer type. The CST value is
1231 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1232
1233 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1234 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1235 pointer type. CST must be of integer type. The CST value is zero extended,
1236 truncated, or unchanged to make it fit in a pointer size. This one is
1237 <i>really</i> dangerous!</dd>
1238
1239 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001240 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1241 identical (same number of bits). The conversion is done as if the CST value
1242 was stored to memory and read back as TYPE. In other words, no bits change
Reid Spencer5c0ef472006-11-11 23:08:07 +00001243 with this operator, just the type. This can be used for conversion of
1244 packed types to any other type, as long as they have the same bit width. For
1245 pointers it is only valid to cast to another pointer type.
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001246 </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001247
1248 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1249
1250 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1251 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1252 instruction, the index list may have zero or more indexes, which are required
1253 to make sense for the type of "CSTPTR".</dd>
1254
Robert Bocchino9fbe1452006-01-10 19:31:34 +00001255 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1256
1257 <dd>Perform the <a href="#i_select">select operation</a> on
1258 constants.
1259
1260 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1261
1262 <dd>Perform the <a href="#i_extractelement">extractelement
1263 operation</a> on constants.
1264
Robert Bocchino05ccd702006-01-15 20:48:27 +00001265 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1266
1267 <dd>Perform the <a href="#i_insertelement">insertelement
1268 operation</a> on constants.
1269
Chris Lattnerc1989542006-04-08 00:13:41 +00001270
1271 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1272
1273 <dd>Perform the <a href="#i_shufflevector">shufflevector
1274 operation</a> on constants.
1275
Chris Lattnerc3f59762004-12-09 17:30:23 +00001276 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1277
Reid Spencer2dc45b82004-12-09 18:13:12 +00001278 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1279 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +00001280 binary</a> operations. The constraints on operands are the same as those for
1281 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +00001282 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001283</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001284</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00001285
Chris Lattner00950542001-06-06 20:29:01 +00001286<!-- *********************************************************************** -->
Chris Lattnere87d6532006-01-25 23:47:57 +00001287<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1288<!-- *********************************************************************** -->
1289
1290<!-- ======================================================================= -->
1291<div class="doc_subsection">
1292<a name="inlineasm">Inline Assembler Expressions</a>
1293</div>
1294
1295<div class="doc_text">
1296
1297<p>
1298LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1299Module-Level Inline Assembly</a>) through the use of a special value. This
1300value represents the inline assembler as a string (containing the instructions
1301to emit), a list of operand constraints (stored as a string), and a flag that
1302indicates whether or not the inline asm expression has side effects. An example
1303inline assembler expression is:
1304</p>
1305
1306<pre>
1307 int(int) asm "bswap $0", "=r,r"
1308</pre>
1309
1310<p>
1311Inline assembler expressions may <b>only</b> be used as the callee operand of
1312a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1313</p>
1314
1315<pre>
1316 %X = call int asm "<a href="#i_bswap">bswap</a> $0", "=r,r"(int %Y)
1317</pre>
1318
1319<p>
1320Inline asms with side effects not visible in the constraint list must be marked
1321as having side effects. This is done through the use of the
1322'<tt>sideeffect</tt>' keyword, like so:
1323</p>
1324
1325<pre>
1326 call void asm sideeffect "eieio", ""()
1327</pre>
1328
1329<p>TODO: The format of the asm and constraints string still need to be
1330documented here. Constraints on what can be done (e.g. duplication, moving, etc
1331need to be documented).
1332</p>
1333
1334</div>
1335
1336<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00001337<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1338<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00001339
Misha Brukman9d0919f2003-11-08 01:05:38 +00001340<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001341
Chris Lattner261efe92003-11-25 01:02:51 +00001342<p>The LLVM instruction set consists of several different
1343classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +00001344instructions</a>, <a href="#binaryops">binary instructions</a>,
1345<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +00001346 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1347instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001348
Misha Brukman9d0919f2003-11-08 01:05:38 +00001349</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001350
Chris Lattner00950542001-06-06 20:29:01 +00001351<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001352<div class="doc_subsection"> <a name="terminators">Terminator
1353Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001354
Misha Brukman9d0919f2003-11-08 01:05:38 +00001355<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001356
Chris Lattner261efe92003-11-25 01:02:51 +00001357<p>As mentioned <a href="#functionstructure">previously</a>, every
1358basic block in a program ends with a "Terminator" instruction, which
1359indicates which block should be executed after the current block is
1360finished. These terminator instructions typically yield a '<tt>void</tt>'
1361value: they produce control flow, not values (the one exception being
1362the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +00001363<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +00001364 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1365instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +00001366the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1367 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1368 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001369
Misha Brukman9d0919f2003-11-08 01:05:38 +00001370</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001371
Chris Lattner00950542001-06-06 20:29:01 +00001372<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001373<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1374Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001375<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001376<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001377<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 +00001378 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001379</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001380<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001381<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +00001382value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +00001383<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +00001384returns a value and then causes control flow, and one that just causes
1385control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001386<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001387<p>The '<tt>ret</tt>' instruction may return any '<a
1388 href="#t_firstclass">first class</a>' type. Notice that a function is
1389not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1390instruction inside of the function that returns a value that does not
1391match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001392<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001393<p>When the '<tt>ret</tt>' instruction is executed, control flow
1394returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +00001395 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +00001396the instruction after the call. If the caller was an "<a
1397 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +00001398at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +00001399returns a value, that value shall set the call or invoke instruction's
1400return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001401<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001402<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001403 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00001404</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001405</div>
Chris Lattner00950542001-06-06 20:29:01 +00001406<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001407<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001408<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001409<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001410<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 +00001411</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001412<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001413<p>The '<tt>br</tt>' instruction is used to cause control flow to
1414transfer to a different basic block in the current function. There are
1415two forms of this instruction, corresponding to a conditional branch
1416and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001417<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001418<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1419single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1420unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1421value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001422<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001423<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1424argument is evaluated. If the value is <tt>true</tt>, control flows
1425to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1426control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001427<h5>Example:</h5>
Reid Spencer36d68e42006-11-18 21:55:45 +00001428<pre>Test:<br> %cond = <a href="#i_icmp">icmp</a> eq, int %a, %b<br> br bool %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
Chris Lattner261efe92003-11-25 01:02:51 +00001429 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 +00001430</div>
Chris Lattner00950542001-06-06 20:29:01 +00001431<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001432<div class="doc_subsubsection">
1433 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1434</div>
1435
Misha Brukman9d0919f2003-11-08 01:05:38 +00001436<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001437<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001438
1439<pre>
1440 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1441</pre>
1442
Chris Lattner00950542001-06-06 20:29:01 +00001443<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001444
1445<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1446several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +00001447instruction, allowing a branch to occur to one of many possible
1448destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001449
1450
Chris Lattner00950542001-06-06 20:29:01 +00001451<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001452
1453<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1454comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1455an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1456table is not allowed to contain duplicate constant entries.</p>
1457
Chris Lattner00950542001-06-06 20:29:01 +00001458<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001459
Chris Lattner261efe92003-11-25 01:02:51 +00001460<p>The <tt>switch</tt> instruction specifies a table of values and
1461destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +00001462table is searched for the given value. If the value is found, control flow is
1463transfered to the corresponding destination; otherwise, control flow is
1464transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001465
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001466<h5>Implementation:</h5>
1467
1468<p>Depending on properties of the target machine and the particular
1469<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +00001470ways. For example, it could be generated as a series of chained conditional
1471branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001472
1473<h5>Example:</h5>
1474
1475<pre>
1476 <i>; Emulate a conditional br instruction</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001477 %Val = <a href="#i_zext">zext</a> bool %value to int
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001478 switch int %Val, label %truedest [int 0, label %falsedest ]
1479
1480 <i>; Emulate an unconditional br instruction</i>
1481 switch uint 0, label %dest [ ]
1482
1483 <i>; Implement a jump table:</i>
1484 switch uint %val, label %otherwise [ uint 0, label %onzero
1485 uint 1, label %onone
1486 uint 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00001487</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001488</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001489
Chris Lattner00950542001-06-06 20:29:01 +00001490<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001491<div class="doc_subsubsection">
1492 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1493</div>
1494
Misha Brukman9d0919f2003-11-08 01:05:38 +00001495<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001496
Chris Lattner00950542001-06-06 20:29:01 +00001497<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001498
1499<pre>
1500 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)
Chris Lattner76b8a332006-05-14 18:23:06 +00001501 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001502</pre>
1503
Chris Lattner6536cfe2002-05-06 22:08:29 +00001504<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001505
1506<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1507function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +00001508'<tt>normal</tt>' label or the
1509'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001510"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1511"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +00001512href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1513continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001514
Chris Lattner00950542001-06-06 20:29:01 +00001515<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001516
Misha Brukman9d0919f2003-11-08 01:05:38 +00001517<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001518
Chris Lattner00950542001-06-06 20:29:01 +00001519<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001520 <li>
John Criswellc1f786c2005-05-13 22:25:59 +00001521 The optional "cconv" marker indicates which <a href="callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001522 convention</a> the call should use. If none is specified, the call defaults
1523 to using C calling conventions.
1524 </li>
1525 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1526 function value being invoked. In most cases, this is a direct function
1527 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1528 an arbitrary pointer to function value.
1529 </li>
1530
1531 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1532 function to be invoked. </li>
1533
1534 <li>'<tt>function args</tt>': argument list whose types match the function
1535 signature argument types. If the function signature indicates the function
1536 accepts a variable number of arguments, the extra arguments can be
1537 specified. </li>
1538
1539 <li>'<tt>normal label</tt>': the label reached when the called function
1540 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1541
1542 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1543 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1544
Chris Lattner00950542001-06-06 20:29:01 +00001545</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001546
Chris Lattner00950542001-06-06 20:29:01 +00001547<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001548
Misha Brukman9d0919f2003-11-08 01:05:38 +00001549<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001550href="#i_call">call</a></tt>' instruction in most regards. The primary
1551difference is that it establishes an association with a label, which is used by
1552the runtime library to unwind the stack.</p>
1553
1554<p>This instruction is used in languages with destructors to ensure that proper
1555cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1556exception. Additionally, this is important for implementation of
1557'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1558
Chris Lattner00950542001-06-06 20:29:01 +00001559<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001560<pre>
1561 %retval = invoke int %Test(int 15) to label %Continue
Chris Lattner76b8a332006-05-14 18:23:06 +00001562 unwind label %TestCleanup <i>; {int}:retval set</i>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001563 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
Chris Lattner76b8a332006-05-14 18:23:06 +00001564 unwind label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00001565</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001566</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001567
1568
Chris Lattner27f71f22003-09-03 00:41:47 +00001569<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00001570
Chris Lattner261efe92003-11-25 01:02:51 +00001571<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1572Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +00001573
Misha Brukman9d0919f2003-11-08 01:05:38 +00001574<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +00001575
Chris Lattner27f71f22003-09-03 00:41:47 +00001576<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001577<pre>
1578 unwind
1579</pre>
1580
Chris Lattner27f71f22003-09-03 00:41:47 +00001581<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001582
1583<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1584at the first callee in the dynamic call stack which used an <a
1585href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1586primarily used to implement exception handling.</p>
1587
Chris Lattner27f71f22003-09-03 00:41:47 +00001588<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00001589
1590<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1591immediately halt. The dynamic call stack is then searched for the first <a
1592href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1593execution continues at the "exceptional" destination block specified by the
1594<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1595dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001596</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001597
1598<!-- _______________________________________________________________________ -->
1599
1600<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1601Instruction</a> </div>
1602
1603<div class="doc_text">
1604
1605<h5>Syntax:</h5>
1606<pre>
1607 unreachable
1608</pre>
1609
1610<h5>Overview:</h5>
1611
1612<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1613instruction is used to inform the optimizer that a particular portion of the
1614code is not reachable. This can be used to indicate that the code after a
1615no-return function cannot be reached, and other facts.</p>
1616
1617<h5>Semantics:</h5>
1618
1619<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1620</div>
1621
1622
1623
Chris Lattner00950542001-06-06 20:29:01 +00001624<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001625<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001626<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001627<p>Binary operators are used to do most of the computation in a
1628program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +00001629produce a single value. The operands might represent
Chris Lattnera58561b2004-08-12 19:12:28 +00001630multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1631The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +00001632necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001633<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001634</div>
Chris Lattner00950542001-06-06 20:29:01 +00001635<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001636<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1637Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001638<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001639<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001640<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 +00001641</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001642<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001643<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001644<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001645<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +00001646 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1647 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1648Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001649<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001650<p>The value produced is the integer or floating point sum of the two
1651operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001652<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001653<pre> &lt;result&gt; = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001654</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001655</div>
Chris Lattner00950542001-06-06 20:29:01 +00001656<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001657<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1658Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001659<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001660<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001661<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 +00001662</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001663<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001664<p>The '<tt>sub</tt>' instruction returns the difference of its two
1665operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001666<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1667instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001668<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001669<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001670 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001671values.
1672This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1673Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001674<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001675<p>The value produced is the integer or floating point difference of
1676the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001677<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001678<pre> &lt;result&gt; = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001679 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1680</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001681</div>
Chris Lattner00950542001-06-06 20:29:01 +00001682<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001683<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1684Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001685<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001686<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001687<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00001688</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001689<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001690<p>The '<tt>mul</tt>' instruction returns the product of its two
1691operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001692<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001693<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00001694 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00001695values.
1696This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1697Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001698<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001699<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +00001700two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001701<p>There is no signed vs unsigned multiplication. The appropriate
1702action is taken based on the type of the operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001703<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001704<pre> &lt;result&gt; = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001705</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001706</div>
Chris Lattner00950542001-06-06 20:29:01 +00001707<!-- _______________________________________________________________________ -->
Reid Spencer1628cec2006-10-26 06:15:43 +00001708<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
1709</a></div>
1710<div class="doc_text">
1711<h5>Syntax:</h5>
1712<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1713</pre>
1714<h5>Overview:</h5>
1715<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
1716operands.</p>
1717<h5>Arguments:</h5>
1718<p>The two arguments to the '<tt>udiv</tt>' instruction must be
1719<a href="#t_integer">integer</a> values. Both arguments must have identical
1720types. This instruction can also take <a href="#t_packed">packed</a> versions
1721of the values in which case the elements must be integers.</p>
1722<h5>Semantics:</h5>
1723<p>The value produced is the unsigned integer quotient of the two operands. This
1724instruction always performs an unsigned division operation, regardless of
1725whether the arguments are unsigned or not.</p>
1726<h5>Example:</h5>
1727<pre> &lt;result&gt; = udiv uint 4, %var <i>; yields {uint}:result = 4 / %var</i>
1728</pre>
1729</div>
1730<!-- _______________________________________________________________________ -->
1731<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
1732</a> </div>
1733<div class="doc_text">
1734<h5>Syntax:</h5>
1735<pre> &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1736</pre>
1737<h5>Overview:</h5>
1738<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
1739operands.</p>
1740<h5>Arguments:</h5>
1741<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
1742<a href="#t_integer">integer</a> values. Both arguments must have identical
1743types. This instruction can also take <a href="#t_packed">packed</a> versions
1744of the values in which case the elements must be integers.</p>
1745<h5>Semantics:</h5>
1746<p>The value produced is the signed integer quotient of the two operands. This
1747instruction always performs a signed division operation, regardless of whether
1748the arguments are signed or not.</p>
1749<h5>Example:</h5>
1750<pre> &lt;result&gt; = sdiv int 4, %var <i>; yields {int}:result = 4 / %var</i>
1751</pre>
1752</div>
1753<!-- _______________________________________________________________________ -->
1754<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00001755Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001756<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001757<h5>Syntax:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00001758<pre> &lt;result&gt; = fdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +00001759</pre>
1760<h5>Overview:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00001761<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
Chris Lattner261efe92003-11-25 01:02:51 +00001762operands.</p>
1763<h5>Arguments:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00001764<p>The two arguments to the '<tt>div</tt>' instruction must be
1765<a href="#t_floating">floating point</a> values. Both arguments must have
1766identical types. This instruction can also take <a href="#t_packed">packed</a>
1767versions of the values in which case the elements must be floating point.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001768<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00001769<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001770<h5>Example:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00001771<pre> &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00001772</pre>
1773</div>
1774<!-- _______________________________________________________________________ -->
Reid Spencer0a783f72006-11-02 01:53:59 +00001775<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
1776</div>
1777<div class="doc_text">
1778<h5>Syntax:</h5>
1779<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1780</pre>
1781<h5>Overview:</h5>
1782<p>The '<tt>urem</tt>' instruction returns the remainder from the
1783unsigned division of its two arguments.</p>
1784<h5>Arguments:</h5>
1785<p>The two arguments to the '<tt>urem</tt>' instruction must be
1786<a href="#t_integer">integer</a> values. Both arguments must have identical
1787types.</p>
1788<h5>Semantics:</h5>
1789<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
1790This instruction always performs an unsigned division to get the remainder,
1791regardless of whether the arguments are unsigned or not.</p>
1792<h5>Example:</h5>
1793<pre> &lt;result&gt; = urem uint 4, %var <i>; yields {uint}:result = 4 % %var</i>
1794</pre>
1795
1796</div>
1797<!-- _______________________________________________________________________ -->
1798<div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00001799Instruction</a> </div>
1800<div class="doc_text">
1801<h5>Syntax:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00001802<pre> &lt;result&gt; = srem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +00001803</pre>
1804<h5>Overview:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00001805<p>The '<tt>srem</tt>' instruction returns the remainder from the
1806signed division of its two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001807<h5>Arguments:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00001808<p>The two arguments to the '<tt>srem</tt>' instruction must be
1809<a href="#t_integer">integer</a> values. Both arguments must have identical
1810types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001811<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00001812<p>This instruction returns the <i>remainder</i> of a division (where the result
Chris Lattner261efe92003-11-25 01:02:51 +00001813has the same sign as the divisor), not the <i>modulus</i> (where the
1814result has the same sign as the dividend) of a value. For more
John Criswell0ec250c2005-10-24 16:17:18 +00001815information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +00001816 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1817Math Forum</a>.</p>
1818<h5>Example:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00001819<pre> &lt;result&gt; = srem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1820</pre>
1821
1822</div>
1823<!-- _______________________________________________________________________ -->
1824<div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
1825Instruction</a> </div>
1826<div class="doc_text">
1827<h5>Syntax:</h5>
1828<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1829</pre>
1830<h5>Overview:</h5>
1831<p>The '<tt>frem</tt>' instruction returns the remainder from the
1832division of its two operands.</p>
1833<h5>Arguments:</h5>
1834<p>The two arguments to the '<tt>frem</tt>' instruction must be
1835<a href="#t_floating">floating point</a> values. Both arguments must have
1836identical types.</p>
1837<h5>Semantics:</h5>
1838<p>This instruction returns the <i>remainder</i> of a division.</p>
1839<h5>Example:</h5>
1840<pre> &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00001841</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001842</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +00001843
Chris Lattner00950542001-06-06 20:29:01 +00001844<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001845<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1846Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001847<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00001848<p>Bitwise binary operators are used to do various forms of
1849bit-twiddling in a program. They are generally very efficient
John Criswell9e2485c2004-12-10 15:51:16 +00001850instructions and can commonly be strength reduced from other
Chris Lattner261efe92003-11-25 01:02:51 +00001851instructions. They require two operands, execute an operation on them,
1852and produce a single value. The resulting value of the bitwise binary
1853operators is always the same type as its first operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001854</div>
Chris Lattner00950542001-06-06 20:29:01 +00001855<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001856<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1857Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001858<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001859<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001860<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 +00001861</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001862<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001863<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1864its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001865<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001866<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001867 href="#t_integral">integral</a> values. Both arguments must have
1868identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001869<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001870<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001871<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001872<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +00001873<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001874 <tbody>
1875 <tr>
1876 <td>In0</td>
1877 <td>In1</td>
1878 <td>Out</td>
1879 </tr>
1880 <tr>
1881 <td>0</td>
1882 <td>0</td>
1883 <td>0</td>
1884 </tr>
1885 <tr>
1886 <td>0</td>
1887 <td>1</td>
1888 <td>0</td>
1889 </tr>
1890 <tr>
1891 <td>1</td>
1892 <td>0</td>
1893 <td>0</td>
1894 </tr>
1895 <tr>
1896 <td>1</td>
1897 <td>1</td>
1898 <td>1</td>
1899 </tr>
1900 </tbody>
1901</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001902</div>
Chris Lattner00950542001-06-06 20:29:01 +00001903<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001904<pre> &lt;result&gt; = and int 4, %var <i>; yields {int}:result = 4 &amp; %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001905 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1906 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1907</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001908</div>
Chris Lattner00950542001-06-06 20:29:01 +00001909<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001910<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001911<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001912<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001913<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 +00001914</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00001915<h5>Overview:</h5>
1916<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1917or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001918<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001919<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001920 href="#t_integral">integral</a> values. Both arguments must have
1921identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001922<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001923<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001924<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001925<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001926<table border="1" cellspacing="0" cellpadding="4">
1927 <tbody>
1928 <tr>
1929 <td>In0</td>
1930 <td>In1</td>
1931 <td>Out</td>
1932 </tr>
1933 <tr>
1934 <td>0</td>
1935 <td>0</td>
1936 <td>0</td>
1937 </tr>
1938 <tr>
1939 <td>0</td>
1940 <td>1</td>
1941 <td>1</td>
1942 </tr>
1943 <tr>
1944 <td>1</td>
1945 <td>0</td>
1946 <td>1</td>
1947 </tr>
1948 <tr>
1949 <td>1</td>
1950 <td>1</td>
1951 <td>1</td>
1952 </tr>
1953 </tbody>
1954</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001955</div>
Chris Lattner00950542001-06-06 20:29:01 +00001956<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001957<pre> &lt;result&gt; = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00001958 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1959 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1960</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001961</div>
Chris Lattner00950542001-06-06 20:29:01 +00001962<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001963<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1964Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001965<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001966<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001967<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 +00001968</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001969<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001970<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1971or of its two operands. The <tt>xor</tt> is used to implement the
1972"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001973<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001974<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +00001975 href="#t_integral">integral</a> values. Both arguments must have
1976identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001977<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001978<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00001979<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001980<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00001981<table border="1" cellspacing="0" cellpadding="4">
1982 <tbody>
1983 <tr>
1984 <td>In0</td>
1985 <td>In1</td>
1986 <td>Out</td>
1987 </tr>
1988 <tr>
1989 <td>0</td>
1990 <td>0</td>
1991 <td>0</td>
1992 </tr>
1993 <tr>
1994 <td>0</td>
1995 <td>1</td>
1996 <td>1</td>
1997 </tr>
1998 <tr>
1999 <td>1</td>
2000 <td>0</td>
2001 <td>1</td>
2002 </tr>
2003 <tr>
2004 <td>1</td>
2005 <td>1</td>
2006 <td>0</td>
2007 </tr>
2008 </tbody>
2009</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00002010</div>
Chris Lattner261efe92003-11-25 01:02:51 +00002011<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +00002012<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002013<pre> &lt;result&gt; = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00002014 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
2015 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner27f71f22003-09-03 00:41:47 +00002016 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00002017</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002018</div>
Chris Lattner00950542001-06-06 20:29:01 +00002019<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002020<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2021Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002022<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002023<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002024<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 +00002025</pre>
Chris Lattner00950542001-06-06 20:29:01 +00002026<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002027<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2028the left a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002029<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002030<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +00002031 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
2032type.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002033<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002034<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002035<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002036<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 +00002037 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
2038 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
2039</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002040</div>
Chris Lattner00950542001-06-06 20:29:01 +00002041<!-- _______________________________________________________________________ -->
Reid Spencer3822ff52006-11-08 06:47:33 +00002042<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00002043Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002044<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002045<h5>Syntax:</h5>
Reid Spencer3822ff52006-11-08 06:47:33 +00002046<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00002047</pre>
Reid Spencer3822ff52006-11-08 06:47:33 +00002048
Chris Lattner00950542001-06-06 20:29:01 +00002049<h5>Overview:</h5>
Reid Spencer3822ff52006-11-08 06:47:33 +00002050<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2051operand shifted to the right a specified number of bits.</p>
2052
Chris Lattner00950542001-06-06 20:29:01 +00002053<h5>Arguments:</h5>
Reid Spencer3822ff52006-11-08 06:47:33 +00002054<p>The first argument to the '<tt>lshr</tt>' instruction must be an <a
2055 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>' type.</p>
2056
Chris Lattner00950542001-06-06 20:29:01 +00002057<h5>Semantics:</h5>
Reid Spencer3822ff52006-11-08 06:47:33 +00002058<p>This instruction always performs a logical shift right operation, regardless
2059of whether the arguments are unsigned or not. The <tt>var2</tt> most significant
2060bits will be filled with zero bits after the shift.</p>
2061
Chris Lattner00950542001-06-06 20:29:01 +00002062<h5>Example:</h5>
Reid Spencer3822ff52006-11-08 06:47:33 +00002063<pre>
2064 &lt;result&gt; = lshr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
2065 &lt;result&gt; = lshr int 4, ubyte 2 <i>; yields {uint}:result = 1</i>
2066 &lt;result&gt; = lshr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
2067 &lt;result&gt; = lshr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = 0x7FFFFFFF </i>
2068</pre>
2069</div>
2070
2071<!-- ======================================================================= -->
2072<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2073Instruction</a> </div>
2074<div class="doc_text">
2075
2076<h5>Syntax:</h5>
2077<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
2078</pre>
2079
2080<h5>Overview:</h5>
2081<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2082operand shifted to the right a specified number of bits.</p>
2083
2084<h5>Arguments:</h5>
2085<p>The first argument to the '<tt>ashr</tt>' instruction must be an
2086<a href="#t_integer">integer</a> type. The second argument must be an
2087'<tt>ubyte</tt>' type.</p>
2088
2089<h5>Semantics:</h5>
2090<p>This instruction always performs an arithmetic shift right operation,
2091regardless of whether the arguments are signed or not. The <tt>var2</tt> most
2092significant bits will be filled with the sign bit of <tt>var1</tt>.</p>
2093
2094<h5>Example:</h5>
2095<pre>
2096 &lt;result&gt; = ashr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
2097 &lt;result&gt; = ashr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
2098 &lt;result&gt; = ashr ubyte 4, ubyte 3 <i>; yields {ubyte}:result = 0</i>
2099 &lt;result&gt; = ashr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner00950542001-06-06 20:29:01 +00002100</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002101</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002102
Chris Lattner00950542001-06-06 20:29:01 +00002103<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002104<div class="doc_subsection">
Chris Lattner3df241e2006-04-08 23:07:04 +00002105 <a name="vectorops">Vector Operations</a>
2106</div>
2107
2108<div class="doc_text">
2109
2110<p>LLVM supports several instructions to represent vector operations in a
2111target-independent manner. This instructions cover the element-access and
2112vector-specific operations needed to process vectors effectively. While LLVM
2113does directly support these vector operations, many sophisticated algorithms
2114will want to use target-specific intrinsics to take full advantage of a specific
2115target.</p>
2116
2117</div>
2118
2119<!-- _______________________________________________________________________ -->
2120<div class="doc_subsubsection">
2121 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2122</div>
2123
2124<div class="doc_text">
2125
2126<h5>Syntax:</h5>
2127
2128<pre>
2129 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, uint &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2130</pre>
2131
2132<h5>Overview:</h5>
2133
2134<p>
2135The '<tt>extractelement</tt>' instruction extracts a single scalar
2136element from a packed vector at a specified index.
2137</p>
2138
2139
2140<h5>Arguments:</h5>
2141
2142<p>
2143The first operand of an '<tt>extractelement</tt>' instruction is a
2144value of <a href="#t_packed">packed</a> type. The second operand is
2145an index indicating the position from which to extract the element.
2146The index may be a variable.</p>
2147
2148<h5>Semantics:</h5>
2149
2150<p>
2151The result is a scalar of the same type as the element type of
2152<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2153<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2154results are undefined.
2155</p>
2156
2157<h5>Example:</h5>
2158
2159<pre>
2160 %result = extractelement &lt;4 x int&gt; %vec, uint 0 <i>; yields int</i>
2161</pre>
2162</div>
2163
2164
2165<!-- _______________________________________________________________________ -->
2166<div class="doc_subsubsection">
2167 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2168</div>
2169
2170<div class="doc_text">
2171
2172<h5>Syntax:</h5>
2173
2174<pre>
2175 &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>
2176</pre>
2177
2178<h5>Overview:</h5>
2179
2180<p>
2181The '<tt>insertelement</tt>' instruction inserts a scalar
2182element into a packed vector at a specified index.
2183</p>
2184
2185
2186<h5>Arguments:</h5>
2187
2188<p>
2189The first operand of an '<tt>insertelement</tt>' instruction is a
2190value of <a href="#t_packed">packed</a> type. The second operand is a
2191scalar value whose type must equal the element type of the first
2192operand. The third operand is an index indicating the position at
2193which to insert the value. The index may be a variable.</p>
2194
2195<h5>Semantics:</h5>
2196
2197<p>
2198The result is a packed vector of the same type as <tt>val</tt>. Its
2199element values are those of <tt>val</tt> except at position
2200<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2201exceeds the length of <tt>val</tt>, the results are undefined.
2202</p>
2203
2204<h5>Example:</h5>
2205
2206<pre>
2207 %result = insertelement &lt;4 x int&gt; %vec, int 1, uint 0 <i>; yields &lt;4 x int&gt;</i>
2208</pre>
2209</div>
2210
2211<!-- _______________________________________________________________________ -->
2212<div class="doc_subsubsection">
2213 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2214</div>
2215
2216<div class="doc_text">
2217
2218<h5>Syntax:</h5>
2219
2220<pre>
2221 &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>
2222</pre>
2223
2224<h5>Overview:</h5>
2225
2226<p>
2227The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2228from two input vectors, returning a vector of the same type.
2229</p>
2230
2231<h5>Arguments:</h5>
2232
2233<p>
2234The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2235with types that match each other and types that match the result of the
2236instruction. The third argument is a shuffle mask, which has the same number
2237of elements as the other vector type, but whose element type is always 'uint'.
2238</p>
2239
2240<p>
2241The shuffle mask operand is required to be a constant vector with either
2242constant integer or undef values.
2243</p>
2244
2245<h5>Semantics:</h5>
2246
2247<p>
2248The elements of the two input vectors are numbered from left to right across
2249both of the vectors. The shuffle mask operand specifies, for each element of
2250the result vector, which element of the two input registers the result element
2251gets. The element selector may be undef (meaning "don't care") and the second
2252operand may be undef if performing a shuffle from only one vector.
2253</p>
2254
2255<h5>Example:</h5>
2256
2257<pre>
2258 %result = shufflevector &lt;4 x int&gt; %v1, &lt;4 x int&gt; %v2,
2259 &lt;4 x uint&gt; &lt;uint 0, uint 4, uint 1, uint 5&gt; <i>; yields &lt;4 x int&gt;</i>
2260 %result = shufflevector &lt;4 x int&gt; %v1, &lt;4 x int&gt; undef,
2261 &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.
2262</pre>
2263</div>
2264
Tanya Lattner09474292006-04-14 19:24:33 +00002265
Chris Lattner3df241e2006-04-08 23:07:04 +00002266<!-- ======================================================================= -->
2267<div class="doc_subsection">
Chris Lattner884a9702006-08-15 00:45:58 +00002268 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002269</div>
2270
Misha Brukman9d0919f2003-11-08 01:05:38 +00002271<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002272
Chris Lattner261efe92003-11-25 01:02:51 +00002273<p>A key design point of an SSA-based representation is how it
2274represents memory. In LLVM, no memory locations are in SSA form, which
2275makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +00002276allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002277
Misha Brukman9d0919f2003-11-08 01:05:38 +00002278</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002279
Chris Lattner00950542001-06-06 20:29:01 +00002280<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002281<div class="doc_subsubsection">
2282 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2283</div>
2284
Misha Brukman9d0919f2003-11-08 01:05:38 +00002285<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002286
Chris Lattner00950542001-06-06 20:29:01 +00002287<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002288
2289<pre>
2290 &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 +00002291</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002292
Chris Lattner00950542001-06-06 20:29:01 +00002293<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002294
Chris Lattner261efe92003-11-25 01:02:51 +00002295<p>The '<tt>malloc</tt>' instruction allocates memory from the system
2296heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002297
Chris Lattner00950542001-06-06 20:29:01 +00002298<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002299
2300<p>The '<tt>malloc</tt>' instruction allocates
2301<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +00002302bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00002303appropriate type to the program. If "NumElements" is specified, it is the
2304number of elements allocated. If an alignment is specified, the value result
2305of the allocation is guaranteed to be aligned to at least that boundary. If
2306not specified, or if zero, the target can choose to align the allocation on any
2307convenient boundary.</p>
2308
Misha Brukman9d0919f2003-11-08 01:05:38 +00002309<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002310
Chris Lattner00950542001-06-06 20:29:01 +00002311<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002312
Chris Lattner261efe92003-11-25 01:02:51 +00002313<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2314a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002315
Chris Lattner2cbdc452005-11-06 08:02:57 +00002316<h5>Example:</h5>
2317
2318<pre>
2319 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
2320
2321 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002322 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
2323 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002324 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
2325 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +00002326</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002327</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002328
Chris Lattner00950542001-06-06 20:29:01 +00002329<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002330<div class="doc_subsubsection">
2331 <a name="i_free">'<tt>free</tt>' Instruction</a>
2332</div>
2333
Misha Brukman9d0919f2003-11-08 01:05:38 +00002334<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002335
Chris Lattner00950542001-06-06 20:29:01 +00002336<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002337
2338<pre>
2339 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +00002340</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002341
Chris Lattner00950542001-06-06 20:29:01 +00002342<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002343
Chris Lattner261efe92003-11-25 01:02:51 +00002344<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +00002345memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002346
Chris Lattner00950542001-06-06 20:29:01 +00002347<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002348
Chris Lattner261efe92003-11-25 01:02:51 +00002349<p>'<tt>value</tt>' shall be a pointer value that points to a value
2350that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2351instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002352
Chris Lattner00950542001-06-06 20:29:01 +00002353<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002354
John Criswell9e2485c2004-12-10 15:51:16 +00002355<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +00002356after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002357
Chris Lattner00950542001-06-06 20:29:01 +00002358<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002359
2360<pre>
2361 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
Chris Lattner00950542001-06-06 20:29:01 +00002362 free [4 x ubyte]* %array
2363</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002364</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002365
Chris Lattner00950542001-06-06 20:29:01 +00002366<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002367<div class="doc_subsubsection">
2368 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2369</div>
2370
Misha Brukman9d0919f2003-11-08 01:05:38 +00002371<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002372
Chris Lattner00950542001-06-06 20:29:01 +00002373<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002374
2375<pre>
2376 &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 +00002377</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002378
Chris Lattner00950542001-06-06 20:29:01 +00002379<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002380
Chris Lattner261efe92003-11-25 01:02:51 +00002381<p>The '<tt>alloca</tt>' instruction allocates memory on the current
2382stack frame of the procedure that is live until the current function
2383returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002384
Chris Lattner00950542001-06-06 20:29:01 +00002385<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002386
John Criswell9e2485c2004-12-10 15:51:16 +00002387<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002388bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00002389appropriate type to the program. If "NumElements" is specified, it is the
2390number of elements allocated. If an alignment is specified, the value result
2391of the allocation is guaranteed to be aligned to at least that boundary. If
2392not specified, or if zero, the target can choose to align the allocation on any
2393convenient boundary.</p>
2394
Misha Brukman9d0919f2003-11-08 01:05:38 +00002395<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002396
Chris Lattner00950542001-06-06 20:29:01 +00002397<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002398
John Criswellc1f786c2005-05-13 22:25:59 +00002399<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +00002400memory is automatically released when the function returns. The '<tt>alloca</tt>'
2401instruction is commonly used to represent automatic variables that must
2402have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +00002403 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002404instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002405
Chris Lattner00950542001-06-06 20:29:01 +00002406<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002407
2408<pre>
2409 %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002410 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002411 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
2412 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00002413</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002414</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002415
Chris Lattner00950542001-06-06 20:29:01 +00002416<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002417<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2418Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002419<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00002420<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002421<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 +00002422<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002423<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002424<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002425<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +00002426address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +00002427 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +00002428marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +00002429the number or order of execution of this <tt>load</tt> with other
2430volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2431instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002432<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002433<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002434<h5>Examples:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002435<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2436 <a
2437 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002438 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2439</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002440</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002441<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002442<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2443Instruction</a> </div>
Reid Spencer035ab572006-11-09 21:18:01 +00002444<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00002445<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002446<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 +00002447 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 +00002448</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002449<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002450<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002451<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002452<p>There are two arguments to the '<tt>store</tt>' instruction: a value
John Criswell0ec250c2005-10-24 16:17:18 +00002453to 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 +00002454operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +00002455operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +00002456optimizer is not allowed to modify the number or order of execution of
2457this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2458 href="#i_store">store</a></tt> instructions.</p>
2459<h5>Semantics:</h5>
2460<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2461at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002462<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002463<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2464 <a
2465 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002466 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2467</pre>
Reid Spencer47ce1792006-11-09 21:15:49 +00002468</div>
2469
Chris Lattner2b7d3202002-05-06 03:03:22 +00002470<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002471<div class="doc_subsubsection">
2472 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2473</div>
2474
Misha Brukman9d0919f2003-11-08 01:05:38 +00002475<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +00002476<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002477<pre>
2478 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2479</pre>
2480
Chris Lattner7faa8832002-04-14 06:13:44 +00002481<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002482
2483<p>
2484The '<tt>getelementptr</tt>' instruction is used to get the address of a
2485subelement of an aggregate data structure.</p>
2486
Chris Lattner7faa8832002-04-14 06:13:44 +00002487<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002488
2489<p>This instruction takes a list of integer constants that indicate what
2490elements of the aggregate object to index to. The actual types of the arguments
2491provided depend on the type of the first pointer argument. The
2492'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +00002493levels of a structure or to a specific index in an array. When indexing into a
Reid Spencer42ddd842006-12-03 16:53:48 +00002494structure, only <tt>uint</tt> integer constants are allowed. When indexing
2495into an array or pointer, integers of any size are allowed, and will be sign
2496extended to 64-bit values.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002497
Chris Lattner261efe92003-11-25 01:02:51 +00002498<p>For example, let's consider a C code fragment and how it gets
2499compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002500
2501<pre>
2502 struct RT {
2503 char A;
2504 int B[10][20];
2505 char C;
2506 };
2507 struct ST {
2508 int X;
2509 double Y;
2510 struct RT Z;
2511 };
2512
2513 int *foo(struct ST *s) {
2514 return &amp;s[1].Z.B[5][13];
2515 }
2516</pre>
2517
Misha Brukman9d0919f2003-11-08 01:05:38 +00002518<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002519
2520<pre>
2521 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2522 %ST = type { int, double, %RT }
2523
Brian Gaeke7283e7c2004-07-02 21:08:14 +00002524 implementation
2525
2526 int* %foo(%ST* %s) {
2527 entry:
2528 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002529 ret int* %reg
2530 }
2531</pre>
2532
Chris Lattner7faa8832002-04-14 06:13:44 +00002533<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002534
2535<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +00002536on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Reid Spencer42ddd842006-12-03 16:53:48 +00002537and <a href="#t_array">array</a> types can use any
2538<a href="#t_integer">integer</a> type but the value will always be sign extended
2539to 64-bits. <a href="#t_struct">Structure</a> types, require <tt>uint</tt>
2540<b>constants</b>.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002541
Misha Brukman9d0919f2003-11-08 01:05:38 +00002542<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002543type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2544}</tt>' type, a structure. The second index indexes into the third element of
2545the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2546sbyte }</tt>' type, another structure. The third index indexes into the second
2547element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2548array. The two dimensions of the array are subscripted into, yielding an
John Criswellfc6b8952005-05-16 16:17:45 +00002549'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002550to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2551
Chris Lattner261efe92003-11-25 01:02:51 +00002552<p>Note that it is perfectly legal to index partially through a
2553structure, returning a pointer to an inner element. Because of this,
2554the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002555
2556<pre>
Chris Lattnerd4f6b172005-03-07 22:13:59 +00002557 int* %foo(%ST* %s) {
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002558 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2559 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2560 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2561 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2562 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2563 ret int* %t5
2564 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00002565</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00002566
2567<p>Note that it is undefined to access an array out of bounds: array and
2568pointer indexes must always be within the defined bounds of the array type.
2569The one exception for this rules is zero length arrays. These arrays are
2570defined to be accessible as variable length arrays, which requires access
2571beyond the zero'th element.</p>
2572
Chris Lattner884a9702006-08-15 00:45:58 +00002573<p>The getelementptr instruction is often confusing. For some more insight
2574into how it works, see <a href="GetElementPtr.html">the getelementptr
2575FAQ</a>.</p>
2576
Chris Lattner7faa8832002-04-14 06:13:44 +00002577<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +00002578
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002579<pre>
2580 <i>; yields [12 x ubyte]*:aptr</i>
2581 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2582</pre>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002583</div>
Reid Spencer47ce1792006-11-09 21:15:49 +00002584
Chris Lattner00950542001-06-06 20:29:01 +00002585<!-- ======================================================================= -->
Reid Spencer2fd21e62006-11-08 01:18:52 +00002586<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002587</div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002588<div class="doc_text">
Reid Spencer2fd21e62006-11-08 01:18:52 +00002589<p>The instructions in this category are the conversion instructions (casting)
2590which all take a single operand and a type. They perform various bit conversions
2591on the operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002592</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002593
Chris Lattner6536cfe2002-05-06 22:08:29 +00002594<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +00002595<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002596 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
2597</div>
2598<div class="doc_text">
2599
2600<h5>Syntax:</h5>
2601<pre>
2602 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2603</pre>
2604
2605<h5>Overview:</h5>
2606<p>
2607The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
2608</p>
2609
2610<h5>Arguments:</h5>
2611<p>
2612The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
2613be an <a href="#t_integer">integer</a> type, and a type that specifies the size
2614and type of the result, which must be an <a href="#t_integral">integral</a>
Reid Spencerd4448792006-11-09 23:03:26 +00002615type. The bit size of <tt>value</tt> must be larger than the bit size of
2616<tt>ty2</tt>. Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002617
2618<h5>Semantics:</h5>
2619<p>
2620The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
Reid Spencerd4448792006-11-09 23:03:26 +00002621and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
2622larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
2623It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002624
2625<h5>Example:</h5>
2626<pre>
2627 %X = trunc int 257 to ubyte <i>; yields ubyte:1</i>
2628 %Y = trunc int 123 to bool <i>; yields bool:true</i>
2629</pre>
2630</div>
2631
2632<!-- _______________________________________________________________________ -->
2633<div class="doc_subsubsection">
2634 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
2635</div>
2636<div class="doc_text">
2637
2638<h5>Syntax:</h5>
2639<pre>
2640 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2641</pre>
2642
2643<h5>Overview:</h5>
2644<p>The '<tt>zext</tt>' instruction zero extends its operand to type
2645<tt>ty2</tt>.</p>
2646
2647
2648<h5>Arguments:</h5>
2649<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
2650<a href="#t_integral">integral</a> type, and a type to cast it to, which must
2651also be of <a href="#t_integral">integral</a> type. The bit size of the
Reid Spencerd4448792006-11-09 23:03:26 +00002652<tt>value</tt> must be smaller than the bit size of the destination type,
2653<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002654
2655<h5>Semantics:</h5>
2656<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
2657bits until it reaches the size of the destination type, <tt>ty2</tt>. When the
2658the operand and the type are the same size, no bit filling is done and the
2659cast is considered a <i>no-op cast</i> because no bits change (only the type
2660changes).</p>
2661
Reid Spencerd4448792006-11-09 23:03:26 +00002662<p>When zero extending from bool, the result will alwasy be either 0 or 1.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002663
2664<h5>Example:</h5>
2665<pre>
2666 %X = zext int 257 to ulong <i>; yields ulong:257</i>
2667 %Y = zext bool true to int <i>; yields int:1</i>
2668</pre>
2669</div>
2670
2671<!-- _______________________________________________________________________ -->
2672<div class="doc_subsubsection">
2673 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
2674</div>
2675<div class="doc_text">
2676
2677<h5>Syntax:</h5>
2678<pre>
2679 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2680</pre>
2681
2682<h5>Overview:</h5>
2683<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
2684
2685<h5>Arguments:</h5>
2686<p>
2687The '<tt>sext</tt>' instruction takes a value to cast, which must be of
2688<a href="#t_integral">integral</a> type, and a type to cast it to, which must
Reid Spencerd4448792006-11-09 23:03:26 +00002689also be of <a href="#t_integral">integral</a> type. The bit size of the
2690<tt>value</tt> must be smaller than the bit size of the destination type,
2691<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002692
2693<h5>Semantics:</h5>
2694<p>
2695The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
2696bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
2697the type <tt>ty2</tt>. When the the operand and the type are the same size,
2698no bit filling is done and the cast is considered a <i>no-op cast</i> because
2699no bits change (only the type changes).</p>
2700
Reid Spencerd4448792006-11-09 23:03:26 +00002701<p>When sign extending from bool, the extension always results in -1 or 0.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002702
2703<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002704<pre>
2705 %X = sext sbyte -1 to ushort <i>; yields ushort:65535</i>
2706 %Y = sext bool true to int <i>; yields int:-1</i>
2707</pre>
2708</div>
2709
2710<!-- _______________________________________________________________________ -->
2711<div class="doc_subsubsection">
Reid Spencer3fa91b02006-11-09 21:48:10 +00002712 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
2713</div>
2714
2715<div class="doc_text">
2716
2717<h5>Syntax:</h5>
2718
2719<pre>
2720 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2721</pre>
2722
2723<h5>Overview:</h5>
2724<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
2725<tt>ty2</tt>.</p>
2726
2727
2728<h5>Arguments:</h5>
2729<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
2730 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
2731cast it to. The size of <tt>value</tt> must be larger than the size of
2732<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
2733<i>no-op cast</i>.</p>
2734
2735<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002736<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
2737<a href="#t_floating">floating point</a> type to a smaller
2738<a href="#t_floating">floating point</a> type. If the value cannot fit within
2739the destination type, <tt>ty2</tt>, then the results are undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00002740
2741<h5>Example:</h5>
2742<pre>
2743 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
2744 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
2745</pre>
2746</div>
2747
2748<!-- _______________________________________________________________________ -->
2749<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002750 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
2751</div>
2752<div class="doc_text">
2753
2754<h5>Syntax:</h5>
2755<pre>
2756 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2757</pre>
2758
2759<h5>Overview:</h5>
2760<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
2761floating point value.</p>
2762
2763<h5>Arguments:</h5>
2764<p>The '<tt>fpext</tt>' instruction takes a
2765<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
Reid Spencerd4448792006-11-09 23:03:26 +00002766and a <a href="#t_floating">floating point</a> type to cast it to. The source
2767type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002768
2769<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002770<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
2771<a href="t_floating">floating point</a> type to a larger
2772<a href="t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
2773used to make a <i>no-op cast</i> because it always changes bits. Use
Reid Spencer5c0ef472006-11-11 23:08:07 +00002774<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002775
2776<h5>Example:</h5>
2777<pre>
2778 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
2779 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
2780</pre>
2781</div>
2782
2783<!-- _______________________________________________________________________ -->
2784<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +00002785 <a name="i_fp2uint">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002786</div>
2787<div class="doc_text">
2788
2789<h5>Syntax:</h5>
2790<pre>
2791 &lt;result&gt; = fp2uint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2792</pre>
2793
2794<h5>Overview:</h5>
2795<p>The '<tt>fp2uint</tt>' converts a floating point <tt>value</tt> to its
2796unsigned integer equivalent of type <tt>ty2</tt>.
2797</p>
2798
2799<h5>Arguments:</h5>
2800<p>The '<tt>fp2uint</tt>' instruction takes a value to cast, which must be a
2801<a href="#t_floating">floating point</a> value, and a type to cast it to, which
2802must be an <a href="#t_integral">integral</a> type.</p>
2803
2804<h5>Semantics:</h5>
2805<p> The '<tt>fp2uint</tt>' instruction converts its
2806<a href="#t_floating">floating point</a> operand into the nearest (rounding
2807towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
2808the results are undefined.</p>
2809
2810<p>When converting to bool, the conversion is done as a comparison against
2811zero. If the <tt>value</tt> was zero, the bool result will be <tt>false</tt>.
2812If the <tt>value</tt> was non-zero, the bool result will be <tt>true</tt>.</p>
2813
2814<h5>Example:</h5>
2815<pre>
2816 %X = fp2uint double 123.0 to int <i>; yields int:123</i>
2817 %Y = fp2uint float 1.0E+300 to bool <i>; yields bool:true</i>
2818 %X = fp2uint float 1.04E+17 to ubyte <i>; yields undefined:1</i>
2819</pre>
2820</div>
2821
2822<!-- _______________________________________________________________________ -->
2823<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +00002824 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002825</div>
2826<div class="doc_text">
2827
2828<h5>Syntax:</h5>
2829<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00002830 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002831</pre>
2832
2833<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002834<p>The '<tt>fptosi</tt>' instruction converts
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002835<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
Chris Lattnercc37aae2004-03-12 05:50:16 +00002836</p>
2837
2838
Chris Lattner6536cfe2002-05-06 22:08:29 +00002839<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002840<p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002841<a href="#t_floating">floating point</a> value, and a type to cast it to, which
2842must also be an <a href="#t_integral">integral</a> type.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002843
Chris Lattner6536cfe2002-05-06 22:08:29 +00002844<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002845<p>The '<tt>fptosi</tt>' instruction converts its
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002846<a href="#t_floating">floating point</a> operand into the nearest (rounding
2847towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
2848the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002849
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002850<p>When converting to bool, the conversion is done as a comparison against
2851zero. If the <tt>value</tt> was zero, the bool result will be <tt>false</tt>.
2852If the <tt>value</tt> was non-zero, the bool result will be <tt>true</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002853
Chris Lattner33ba0d92001-07-09 00:26:23 +00002854<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00002855<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00002856 %X = fptosi double -123.0 to int <i>; yields int:-123</i>
2857 %Y = fptosi float 1.0E-247 to bool <i>; yields bool:true</i>
2858 %X = fptosi float 1.04E+17 to sbyte <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002859</pre>
2860</div>
2861
2862<!-- _______________________________________________________________________ -->
2863<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +00002864 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002865</div>
2866<div class="doc_text">
2867
2868<h5>Syntax:</h5>
2869<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00002870 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002871</pre>
2872
2873<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002874<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002875integer and converts that value to the <tt>ty2</tt> type.</p>
2876
2877
2878<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002879<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be an
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002880<a href="#t_integral">integral</a> value, and a type to cast it to, which must
2881be a <a href="#t_floating">floating point</a> type.</p>
2882
2883<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002884<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002885integer quantity and converts it to the corresponding floating point value. If
2886the value cannot fit in the floating point value, the results are undefined.</p>
2887
2888
2889<h5>Example:</h5>
2890<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00002891 %X = uitofp int 257 to float <i>; yields float:257.0</i>
2892 %Y = uitofp sbyte -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002893</pre>
2894</div>
2895
2896<!-- _______________________________________________________________________ -->
2897<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +00002898 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002899</div>
2900<div class="doc_text">
2901
2902<h5>Syntax:</h5>
2903<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00002904 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002905</pre>
2906
2907<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002908<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002909integer and converts that value to the <tt>ty2</tt> type.</p>
2910
2911<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002912<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be an
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002913<a href="#t_integral">integral</a> value, and a type to cast it to, which must be
2914a <a href="#t_floating">floating point</a> type.</p>
2915
2916<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00002917<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002918integer quantity and converts it to the corresponding floating point value. If
2919the value cannot fit in the floating point value, the results are undefined.</p>
2920
2921<h5>Example:</h5>
2922<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00002923 %X = sitofp int 257 to float <i>; yields float:257.0</i>
2924 %Y = sitofp sbyte -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002925</pre>
2926</div>
2927
2928<!-- _______________________________________________________________________ -->
2929<div class="doc_subsubsection">
Reid Spencer72679252006-11-11 21:00:47 +00002930 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
2931</div>
2932<div class="doc_text">
2933
2934<h5>Syntax:</h5>
2935<pre>
2936 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2937</pre>
2938
2939<h5>Overview:</h5>
2940<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
2941the integer type <tt>ty2</tt>.</p>
2942
2943<h5>Arguments:</h5>
2944<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
2945must be a <a href="t_pointer">pointer</a> value, and a type to cast it to
2946<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
2947
2948<h5>Semantics:</h5>
2949<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
2950<tt>ty2</tt> by interpreting the pointer value as an integer and either
2951truncating or zero extending that value to the size of the integer type. If
2952<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
2953<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
2954are the same size, then nothing is done (<i>no-op cast</i>).</p>
2955
2956<h5>Example:</h5>
2957<pre>
2958 %X = ptrtoint int* %X to sbyte <i>; yields truncation on 32-bit</i>
2959 %Y = ptrtoint int* %x to ulong <i>; yields zero extend on 32-bit</i>
2960</pre>
2961</div>
2962
2963<!-- _______________________________________________________________________ -->
2964<div class="doc_subsubsection">
2965 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
2966</div>
2967<div class="doc_text">
2968
2969<h5>Syntax:</h5>
2970<pre>
2971 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2972</pre>
2973
2974<h5>Overview:</h5>
2975<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
2976a pointer type, <tt>ty2</tt>.</p>
2977
2978<h5>Arguments:</h5>
2979<p>The '<tt>inttoptr</tt>' instruction takes an <a href="i_integer">integer</a>
2980value to cast, and a type to cast it to, which must be a
2981<a href="#t_pointer">pointer</a> type. </tt>
2982
2983<h5>Semantics:</h5>
2984<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
2985<tt>ty2</tt> by applying either a zero extension or a truncation depending on
2986the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
2987size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
2988the size of a pointer then a zero extension is done. If they are the same size,
2989nothing is done (<i>no-op cast</i>).</p>
2990
2991<h5>Example:</h5>
2992<pre>
2993 %X = inttoptr int 255 to int* <i>; yields zero extend on 64-bit</i>
2994 %X = inttoptr int 255 to int* <i>; yields no-op on 32-bit </i>
2995 %Y = inttoptr short 0 to int* <i>; yields zero extend on 32-bit</i>
2996</pre>
2997</div>
2998
2999<!-- _______________________________________________________________________ -->
3000<div class="doc_subsubsection">
Reid Spencer5c0ef472006-11-11 23:08:07 +00003001 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003002</div>
3003<div class="doc_text">
3004
3005<h5>Syntax:</h5>
3006<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003007 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003008</pre>
3009
3010<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003011<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003012<tt>ty2</tt> without changing any bits.</p>
3013
3014<h5>Arguments:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003015<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003016a first class value, and a type to cast it to, which must also be a <a
3017 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
3018and the destination type, <tt>ty2</tt>, must be identical.</p>
3019
3020<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003021<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer72679252006-11-11 21:00:47 +00003022<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3023this conversion. The conversion is done as if the <tt>value</tt> had been
3024stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3025converted to other pointer types with this instruction. To convert pointers to
3026other types, use the <a href="#i_inttoptr">inttoptr</a> or
3027<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003028
3029<h5>Example:</h5>
3030<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003031 %X = bitcast ubyte 255 to sbyte <i>; yields sbyte:-1</i>
3032 %Y = bitcast uint* %x to sint* <i>; yields sint*:%x</i>
3033 %Z = bitcast <2xint> %V to long; <i>; yields long: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00003034</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003035</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003036
Reid Spencer2fd21e62006-11-08 01:18:52 +00003037<!-- ======================================================================= -->
3038<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3039<div class="doc_text">
3040<p>The instructions in this category are the "miscellaneous"
3041instructions, which defy better classification.</p>
3042</div>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003043
3044<!-- _______________________________________________________________________ -->
3045<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3046</div>
3047<div class="doc_text">
3048<h5>Syntax:</h5>
3049<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
3050</pre>
3051<h5>Overview:</h5>
3052<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
3053of its two integer operands.</p>
3054<h5>Arguments:</h5>
3055<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3056the condition code which indicates the kind of comparison to perform. It is not
3057a value, just a keyword. The possibilities for the condition code are:
3058<ol>
3059 <li><tt>eq</tt>: equal</li>
3060 <li><tt>ne</tt>: not equal </li>
3061 <li><tt>ugt</tt>: unsigned greater than</li>
3062 <li><tt>uge</tt>: unsigned greater or equal</li>
3063 <li><tt>ult</tt>: unsigned less than</li>
3064 <li><tt>ule</tt>: unsigned less or equal</li>
3065 <li><tt>sgt</tt>: signed greater than</li>
3066 <li><tt>sge</tt>: signed greater or equal</li>
3067 <li><tt>slt</tt>: signed less than</li>
3068 <li><tt>sle</tt>: signed less or equal</li>
3069</ol>
3070<p>The remaining two arguments must be of <a href="#t_integral">integral</a>,
3071<a href="#t_pointer">pointer</a> or a <a href="#t_packed">packed</a> integral
3072type. They must have identical types.</p>
3073<h5>Semantics:</h5>
3074<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3075the condition code given as <tt>cond</tt>. The comparison performed always
3076yields a <a href="#t_bool">bool</a> result, as follows:
3077<ol>
3078 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3079 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3080 </li>
3081 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3082 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3083 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3084 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3085 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3086 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3087 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3088 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3089 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3090 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3091 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3092 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3093 <li><tt>sge</tt>: interprets the operands as signed values and yields
3094 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3095 <li><tt>slt</tt>: interprets the operands as signed values and yields
3096 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3097 <li><tt>sle</tt>: interprets the operands as signed values and yields
3098 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3099 </li>
3100</ol>
3101<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3102values are treated as integers and then compared.</p>
3103<p>If the operands are <a href="#t_packed">packed</a> typed, the elements of
Reid Spencerb7f26282006-11-19 03:00:14 +00003104the vector are compared in turn and the predicate must hold for all
3105elements.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003106
3107<h5>Example:</h5>
3108<pre> &lt;result&gt; = icmp eq int 4, 5 <i>; yields: result=false</i>
3109 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3110 &lt;result&gt; = icmp ult short 4, 5 <i>; yields: result=true</i>
3111 &lt;result&gt; = icmp sgt sbyte 4, 5 <i>; yields: result=false</i>
3112 &lt;result&gt; = icmp ule sbyte -4, 5 <i>; yields: result=false</i>
3113 &lt;result&gt; = icmp sge sbyte 4, 5 <i>; yields: result=false</i>
3114</pre>
3115</div>
3116
3117<!-- _______________________________________________________________________ -->
3118<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3119</div>
3120<div class="doc_text">
3121<h5>Syntax:</h5>
3122<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
3123</pre>
3124<h5>Overview:</h5>
3125<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3126of its floating point operands.</p>
3127<h5>Arguments:</h5>
3128<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3129the condition code which indicates the kind of comparison to perform. It is not
3130a value, just a keyword. The possibilities for the condition code are:
3131<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +00003132 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003133 <li><tt>oeq</tt>: ordered and equal</li>
3134 <li><tt>ogt</tt>: ordered and greater than </li>
3135 <li><tt>oge</tt>: ordered and greater than or equal</li>
3136 <li><tt>olt</tt>: ordered and less than </li>
3137 <li><tt>ole</tt>: ordered and less than or equal</li>
3138 <li><tt>one</tt>: ordered and not equal</li>
3139 <li><tt>ord</tt>: ordered (no nans)</li>
3140 <li><tt>ueq</tt>: unordered or equal</li>
3141 <li><tt>ugt</tt>: unordered or greater than </li>
3142 <li><tt>uge</tt>: unordered or greater than or equal</li>
3143 <li><tt>ult</tt>: unordered or less than </li>
3144 <li><tt>ule</tt>: unordered or less than or equal</li>
3145 <li><tt>une</tt>: unordered or not equal</li>
3146 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003147 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003148</ol>
3149<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be of
3150<a href="#t_floating">floating point</a>, or a <a href="#t_packed">packed</a>
3151floating point type. They must have identical types.</p>
Reid Spencerb7f26282006-11-19 03:00:14 +00003152<p>In the foregoing, <i>ordered</i> means that neither operand is a QNAN and
3153<i>unordered</i> means that either operand is a QNAN.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003154<h5>Semantics:</h5>
3155<p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3156the condition code given as <tt>cond</tt>. The comparison performed always
3157yields a <a href="#t_bool">bool</a> result, as follows:
3158<ol>
3159 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003160 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003161 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003162 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003163 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003164 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003165 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003166 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003167 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003168 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003169 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003170 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003171 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003172 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3173 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003174 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003175 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003176 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003177 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003178 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003179 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003180 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003181 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003182 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003183 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003184 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003185 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003186 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3187</ol>
3188<p>If the operands are <a href="#t_packed">packed</a> typed, the elements of
3189the vector are compared in turn and the predicate must hold for all elements.
Reid Spencerb7f26282006-11-19 03:00:14 +00003190</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003191
3192<h5>Example:</h5>
3193<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3194 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3195 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3196 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3197</pre>
3198</div>
3199
Reid Spencer2fd21e62006-11-08 01:18:52 +00003200<!-- _______________________________________________________________________ -->
3201<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3202Instruction</a> </div>
3203<div class="doc_text">
3204<h5>Syntax:</h5>
3205<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
3206<h5>Overview:</h5>
3207<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
3208the SSA graph representing the function.</p>
3209<h5>Arguments:</h5>
3210<p>The type of the incoming values are specified with the first type
3211field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3212as arguments, with one pair for each predecessor basic block of the
3213current block. Only values of <a href="#t_firstclass">first class</a>
3214type may be used as the value arguments to the PHI node. Only labels
3215may be used as the label arguments.</p>
3216<p>There must be no non-phi instructions between the start of a basic
3217block and the PHI instructions: i.e. PHI instructions must be first in
3218a basic block.</p>
3219<h5>Semantics:</h5>
3220<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
3221value specified by the parameter, depending on which basic block we
3222came from in the last <a href="#terminators">terminator</a> instruction.</p>
3223<h5>Example:</h5>
3224<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>
3225</div>
3226
Chris Lattnercc37aae2004-03-12 05:50:16 +00003227<!-- _______________________________________________________________________ -->
3228<div class="doc_subsubsection">
3229 <a name="i_select">'<tt>select</tt>' Instruction</a>
3230</div>
3231
3232<div class="doc_text">
3233
3234<h5>Syntax:</h5>
3235
3236<pre>
3237 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
3238</pre>
3239
3240<h5>Overview:</h5>
3241
3242<p>
3243The '<tt>select</tt>' instruction is used to choose one value based on a
3244condition, without branching.
3245</p>
3246
3247
3248<h5>Arguments:</h5>
3249
3250<p>
3251The '<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.
3252</p>
3253
3254<h5>Semantics:</h5>
3255
3256<p>
3257If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +00003258value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +00003259</p>
3260
3261<h5>Example:</h5>
3262
3263<pre>
3264 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
3265</pre>
3266</div>
3267
Robert Bocchino05ccd702006-01-15 20:48:27 +00003268
3269<!-- _______________________________________________________________________ -->
3270<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +00003271 <a name="i_call">'<tt>call</tt>' Instruction</a>
3272</div>
3273
Misha Brukman9d0919f2003-11-08 01:05:38 +00003274<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +00003275
Chris Lattner00950542001-06-06 20:29:01 +00003276<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003277<pre>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003278 &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 +00003279</pre>
3280
Chris Lattner00950542001-06-06 20:29:01 +00003281<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003282
Misha Brukman9d0919f2003-11-08 01:05:38 +00003283<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003284
Chris Lattner00950542001-06-06 20:29:01 +00003285<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003286
Misha Brukman9d0919f2003-11-08 01:05:38 +00003287<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003288
Chris Lattner6536cfe2002-05-06 22:08:29 +00003289<ol>
Chris Lattner261efe92003-11-25 01:02:51 +00003290 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003291 <p>The optional "tail" marker indicates whether the callee function accesses
3292 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +00003293 function call is eligible for tail call optimization. Note that calls may
3294 be marked "tail" even if they do not occur before a <a
3295 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +00003296 </li>
3297 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003298 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
3299 convention</a> the call should use. If none is specified, the call defaults
3300 to using C calling conventions.
3301 </li>
3302 <li>
Chris Lattner2bff5242005-05-06 05:47:36 +00003303 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
3304 being invoked. The argument types must match the types implied by this
John Criswellfc6b8952005-05-16 16:17:45 +00003305 signature. This type can be omitted if the function is not varargs and
3306 if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003307 </li>
3308 <li>
3309 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3310 be invoked. In most cases, this is a direct function invocation, but
3311 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +00003312 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00003313 </li>
3314 <li>
3315 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +00003316 function signature argument types. All arguments must be of
3317 <a href="#t_firstclass">first class</a> type. If the function signature
3318 indicates the function accepts a variable number of arguments, the extra
3319 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00003320 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00003321</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00003322
Chris Lattner00950542001-06-06 20:29:01 +00003323<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003324
Chris Lattner261efe92003-11-25 01:02:51 +00003325<p>The '<tt>call</tt>' instruction is used to cause control flow to
3326transfer to a specified function, with its incoming arguments bound to
3327the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3328instruction in the called function, control flow continues with the
3329instruction after the function call, and the return value of the
3330function is bound to the result argument. This is a simpler case of
3331the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003332
Chris Lattner00950542001-06-06 20:29:01 +00003333<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003334
3335<pre>
3336 %retval = call int %test(int %argc)
3337 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
3338 %X = tail call int %foo()
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003339 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
Chris Lattner2bff5242005-05-06 05:47:36 +00003340</pre>
3341
Misha Brukman9d0919f2003-11-08 01:05:38 +00003342</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003343
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003344<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +00003345<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +00003346 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003347</div>
3348
Misha Brukman9d0919f2003-11-08 01:05:38 +00003349<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +00003350
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003351<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003352
3353<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003354 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00003355</pre>
3356
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003357<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003358
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003359<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +00003360the "variable argument" area of a function call. It is used to implement the
3361<tt>va_arg</tt> macro in C.</p>
3362
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003363<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003364
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003365<p>This instruction takes a <tt>va_list*</tt> value and the type of
3366the argument. It returns a value of the specified argument type and
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00003367increments the <tt>va_list</tt> to point to the next argument. Again, the
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003368actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003369
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003370<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003371
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003372<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3373type from the specified <tt>va_list</tt> and causes the
3374<tt>va_list</tt> to point to the next argument. For more information,
3375see the variable argument handling <a href="#int_varargs">Intrinsic
3376Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003377
3378<p>It is legal for this instruction to be called in a function which does not
3379take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003380function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003381
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003382<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +00003383href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +00003384argument.</p>
3385
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003386<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003387
3388<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3389
Misha Brukman9d0919f2003-11-08 01:05:38 +00003390</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003391
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003392<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00003393<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3394<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00003395
Misha Brukman9d0919f2003-11-08 01:05:38 +00003396<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +00003397
3398<p>LLVM supports the notion of an "intrinsic function". These functions have
John Criswellfc6b8952005-05-16 16:17:45 +00003399well known names and semantics and are required to follow certain
Chris Lattner33aec9e2004-02-12 17:01:32 +00003400restrictions. Overall, these instructions represent an extension mechanism for
3401the LLVM language that does not require changing all of the transformations in
3402LLVM to add to the language (or the bytecode reader/writer, the parser,
3403etc...).</p>
3404
John Criswellfc6b8952005-05-16 16:17:45 +00003405<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
3406prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
Chris Lattner33aec9e2004-02-12 17:01:32 +00003407this. Intrinsic functions must always be external functions: you cannot define
3408the body of intrinsic functions. Intrinsic functions may only be used in call
3409or invoke instructions: it is illegal to take the address of an intrinsic
3410function. Additionally, because intrinsic functions are part of the LLVM
3411language, it is required that they all be documented here if any are added.</p>
3412
3413
John Criswellfc6b8952005-05-16 16:17:45 +00003414<p>To learn how to add an intrinsic function, please see the <a
Chris Lattner590cff32005-05-11 03:35:57 +00003415href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003416</p>
3417
Misha Brukman9d0919f2003-11-08 01:05:38 +00003418</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003419
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003420<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +00003421<div class="doc_subsection">
3422 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
3423</div>
3424
Misha Brukman9d0919f2003-11-08 01:05:38 +00003425<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00003426
Misha Brukman9d0919f2003-11-08 01:05:38 +00003427<p>Variable argument support is defined in LLVM with the <a
Chris Lattnerfb6977d2006-01-13 23:26:01 +00003428 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner261efe92003-11-25 01:02:51 +00003429intrinsic functions. These functions are related to the similarly
3430named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00003431
Chris Lattner261efe92003-11-25 01:02:51 +00003432<p>All of these functions operate on arguments that use a
3433target-specific value type "<tt>va_list</tt>". The LLVM assembly
3434language reference manual does not define what this type is, so all
3435transformations should be prepared to handle intrinsics with any type
3436used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00003437
Chris Lattner374ab302006-05-15 17:26:46 +00003438<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +00003439instruction and the variable argument handling intrinsic functions are
3440used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00003441
Chris Lattner33aec9e2004-02-12 17:01:32 +00003442<pre>
3443int %test(int %X, ...) {
3444 ; Initialize variable argument processing
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003445 %ap = alloca sbyte*
3446 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003447
3448 ; Read a single integer argument
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003449 %tmp = va_arg sbyte** %ap, int
Chris Lattner33aec9e2004-02-12 17:01:32 +00003450
3451 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003452 %aq = alloca sbyte*
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00003453 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003454 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003455
3456 ; Stop processing of arguments.
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003457 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003458 ret int %tmp
3459}
3460</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003461</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003462
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003463<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00003464<div class="doc_subsubsection">
3465 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
3466</div>
3467
3468
Misha Brukman9d0919f2003-11-08 01:05:38 +00003469<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003470<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003471<pre> declare void %llvm.va_start(&lt;va_list&gt;* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003472<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003473<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
3474<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
3475href="#i_va_arg">va_arg</a></tt>.</p>
3476
3477<h5>Arguments:</h5>
3478
3479<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
3480
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003481<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003482
3483<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
3484macro available in C. In a target-dependent way, it initializes the
3485<tt>va_list</tt> element the argument points to, so that the next call to
3486<tt>va_arg</tt> will produce the first variable argument passed to the function.
3487Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
3488last argument of the function, the compiler can figure that out.</p>
3489
Misha Brukman9d0919f2003-11-08 01:05:38 +00003490</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003491
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003492<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00003493<div class="doc_subsubsection">
3494 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
3495</div>
3496
Misha Brukman9d0919f2003-11-08 01:05:38 +00003497<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003498<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003499<pre> declare void %llvm.va_end(&lt;va_list*&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003500<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00003501<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
3502which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
3503or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003504<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003505<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003506<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003507<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00003508macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
3509Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
3510 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
3511with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003512</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003513
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003514<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00003515<div class="doc_subsubsection">
3516 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
3517</div>
3518
Misha Brukman9d0919f2003-11-08 01:05:38 +00003519<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00003520
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003521<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00003522
3523<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003524 declare void %llvm.va_copy(&lt;va_list&gt;* &lt;destarglist&gt;,
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00003525 &lt;va_list&gt;* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00003526</pre>
3527
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003528<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00003529
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003530<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
3531the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00003532
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003533<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00003534
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003535<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00003536The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003537
Chris Lattnerd7923912004-05-23 21:06:01 +00003538
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003539<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00003540
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003541<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
3542available in C. In a target-dependent way, it copies the source
3543<tt>va_list</tt> element into the destination list. This intrinsic is necessary
3544because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
Chris Lattnerd7923912004-05-23 21:06:01 +00003545arbitrarily complex and require memory allocation, for example.</p>
3546
Misha Brukman9d0919f2003-11-08 01:05:38 +00003547</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003548
Chris Lattner33aec9e2004-02-12 17:01:32 +00003549<!-- ======================================================================= -->
3550<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +00003551 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
3552</div>
3553
3554<div class="doc_text">
3555
3556<p>
3557LLVM support for <a href="GarbageCollection.html">Accurate Garbage
3558Collection</a> requires the implementation and generation of these intrinsics.
3559These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
3560stack</a>, as well as garbage collector implementations that require <a
3561href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
3562Front-ends for type-safe garbage collected languages should generate these
3563intrinsics to make use of the LLVM garbage collectors. For more details, see <a
3564href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
3565</p>
3566</div>
3567
3568<!-- _______________________________________________________________________ -->
3569<div class="doc_subsubsection">
3570 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
3571</div>
3572
3573<div class="doc_text">
3574
3575<h5>Syntax:</h5>
3576
3577<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003578 declare void %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00003579</pre>
3580
3581<h5>Overview:</h5>
3582
John Criswell9e2485c2004-12-10 15:51:16 +00003583<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +00003584the code generator, and allows some metadata to be associated with it.</p>
3585
3586<h5>Arguments:</h5>
3587
3588<p>The first argument specifies the address of a stack object that contains the
3589root pointer. The second pointer (which must be either a constant or a global
3590value address) contains the meta-data to be associated with the root.</p>
3591
3592<h5>Semantics:</h5>
3593
3594<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
3595location. At compile-time, the code generator generates information to allow
3596the runtime to find the pointer at GC safe points.
3597</p>
3598
3599</div>
3600
3601
3602<!-- _______________________________________________________________________ -->
3603<div class="doc_subsubsection">
3604 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
3605</div>
3606
3607<div class="doc_text">
3608
3609<h5>Syntax:</h5>
3610
3611<pre>
Chris Lattner80626e92006-03-14 20:02:51 +00003612 declare sbyte* %llvm.gcread(sbyte* %ObjPtr, sbyte** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00003613</pre>
3614
3615<h5>Overview:</h5>
3616
3617<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
3618locations, allowing garbage collector implementations that require read
3619barriers.</p>
3620
3621<h5>Arguments:</h5>
3622
Chris Lattner80626e92006-03-14 20:02:51 +00003623<p>The second argument is the address to read from, which should be an address
3624allocated from the garbage collector. The first object is a pointer to the
3625start of the referenced object, if needed by the language runtime (otherwise
3626null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00003627
3628<h5>Semantics:</h5>
3629
3630<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
3631instruction, but may be replaced with substantially more complex code by the
3632garbage collector runtime, as needed.</p>
3633
3634</div>
3635
3636
3637<!-- _______________________________________________________________________ -->
3638<div class="doc_subsubsection">
3639 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
3640</div>
3641
3642<div class="doc_text">
3643
3644<h5>Syntax:</h5>
3645
3646<pre>
Chris Lattner80626e92006-03-14 20:02:51 +00003647 declare void %llvm.gcwrite(sbyte* %P1, sbyte* %Obj, sbyte** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00003648</pre>
3649
3650<h5>Overview:</h5>
3651
3652<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
3653locations, allowing garbage collector implementations that require write
3654barriers (such as generational or reference counting collectors).</p>
3655
3656<h5>Arguments:</h5>
3657
Chris Lattner80626e92006-03-14 20:02:51 +00003658<p>The first argument is the reference to store, the second is the start of the
3659object to store it to, and the third is the address of the field of Obj to
3660store to. If the runtime does not require a pointer to the object, Obj may be
3661null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00003662
3663<h5>Semantics:</h5>
3664
3665<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
3666instruction, but may be replaced with substantially more complex code by the
3667garbage collector runtime, as needed.</p>
3668
3669</div>
3670
3671
3672
3673<!-- ======================================================================= -->
3674<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +00003675 <a name="int_codegen">Code Generator Intrinsics</a>
3676</div>
3677
3678<div class="doc_text">
3679<p>
3680These intrinsics are provided by LLVM to expose special features that may only
3681be implemented with code generator support.
3682</p>
3683
3684</div>
3685
3686<!-- _______________________________________________________________________ -->
3687<div class="doc_subsubsection">
3688 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
3689</div>
3690
3691<div class="doc_text">
3692
3693<h5>Syntax:</h5>
3694<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00003695 declare sbyte *%llvm.returnaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003696</pre>
3697
3698<h5>Overview:</h5>
3699
3700<p>
Chris Lattner32b5d712006-10-15 20:05:59 +00003701The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
3702target-specific value indicating the return address of the current function
3703or one of its callers.
Chris Lattner10610642004-02-14 04:08:35 +00003704</p>
3705
3706<h5>Arguments:</h5>
3707
3708<p>
3709The argument to this intrinsic indicates which function to return the address
3710for. Zero indicates the calling function, one indicates its caller, etc. The
3711argument is <b>required</b> to be a constant integer value.
3712</p>
3713
3714<h5>Semantics:</h5>
3715
3716<p>
3717The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
3718the return address of the specified call frame, or zero if it cannot be
3719identified. The value returned by this intrinsic is likely to be incorrect or 0
3720for arguments other than zero, so it should only be used for debugging purposes.
3721</p>
3722
3723<p>
3724Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00003725aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00003726source-language caller.
3727</p>
3728</div>
3729
3730
3731<!-- _______________________________________________________________________ -->
3732<div class="doc_subsubsection">
3733 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
3734</div>
3735
3736<div class="doc_text">
3737
3738<h5>Syntax:</h5>
3739<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +00003740 declare sbyte *%llvm.frameaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00003741</pre>
3742
3743<h5>Overview:</h5>
3744
3745<p>
Chris Lattner32b5d712006-10-15 20:05:59 +00003746The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
3747target-specific frame pointer value for the specified stack frame.
Chris Lattner10610642004-02-14 04:08:35 +00003748</p>
3749
3750<h5>Arguments:</h5>
3751
3752<p>
3753The argument to this intrinsic indicates which function to return the frame
3754pointer for. Zero indicates the calling function, one indicates its caller,
3755etc. The argument is <b>required</b> to be a constant integer value.
3756</p>
3757
3758<h5>Semantics:</h5>
3759
3760<p>
3761The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
3762the frame address of the specified call frame, or zero if it cannot be
3763identified. The value returned by this intrinsic is likely to be incorrect or 0
3764for arguments other than zero, so it should only be used for debugging purposes.
3765</p>
3766
3767<p>
3768Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00003769aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00003770source-language caller.
3771</p>
3772</div>
3773
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003774<!-- _______________________________________________________________________ -->
3775<div class="doc_subsubsection">
Chris Lattner57e1f392006-01-13 02:03:13 +00003776 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
3777</div>
3778
3779<div class="doc_text">
3780
3781<h5>Syntax:</h5>
3782<pre>
3783 declare sbyte *%llvm.stacksave()
3784</pre>
3785
3786<h5>Overview:</h5>
3787
3788<p>
3789The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
3790the function stack, for use with <a href="#i_stackrestore">
3791<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
3792features like scoped automatic variable sized arrays in C99.
3793</p>
3794
3795<h5>Semantics:</h5>
3796
3797<p>
3798This intrinsic returns a opaque pointer value that can be passed to <a
3799href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
3800<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
3801<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
3802state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
3803practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
3804that were allocated after the <tt>llvm.stacksave</tt> was executed.
3805</p>
3806
3807</div>
3808
3809<!-- _______________________________________________________________________ -->
3810<div class="doc_subsubsection">
3811 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
3812</div>
3813
3814<div class="doc_text">
3815
3816<h5>Syntax:</h5>
3817<pre>
3818 declare void %llvm.stackrestore(sbyte* %ptr)
3819</pre>
3820
3821<h5>Overview:</h5>
3822
3823<p>
3824The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
3825the function stack to the state it was in when the corresponding <a
3826href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
3827useful for implementing language features like scoped automatic variable sized
3828arrays in C99.
3829</p>
3830
3831<h5>Semantics:</h5>
3832
3833<p>
3834See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
3835</p>
3836
3837</div>
3838
3839
3840<!-- _______________________________________________________________________ -->
3841<div class="doc_subsubsection">
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003842 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
3843</div>
3844
3845<div class="doc_text">
3846
3847<h5>Syntax:</h5>
3848<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003849 declare void %llvm.prefetch(sbyte * &lt;address&gt;,
3850 uint &lt;rw&gt;, uint &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003851</pre>
3852
3853<h5>Overview:</h5>
3854
3855
3856<p>
3857The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +00003858a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
3859no
3860effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +00003861characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003862</p>
3863
3864<h5>Arguments:</h5>
3865
3866<p>
3867<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
3868determining if the fetch should be for a read (0) or write (1), and
3869<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +00003870locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00003871<tt>locality</tt> arguments must be constant integers.
3872</p>
3873
3874<h5>Semantics:</h5>
3875
3876<p>
3877This intrinsic does not modify the behavior of the program. In particular,
3878prefetches cannot trap and do not produce a value. On targets that support this
3879intrinsic, the prefetch can provide hints to the processor cache for better
3880performance.
3881</p>
3882
3883</div>
3884
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003885<!-- _______________________________________________________________________ -->
3886<div class="doc_subsubsection">
3887 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
3888</div>
3889
3890<div class="doc_text">
3891
3892<h5>Syntax:</h5>
3893<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +00003894 declare void %llvm.pcmarker( uint &lt;id&gt; )
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003895</pre>
3896
3897<h5>Overview:</h5>
3898
3899
3900<p>
John Criswellfc6b8952005-05-16 16:17:45 +00003901The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
3902(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003903code to simulators and other tools. The method is target specific, but it is
3904expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00003905The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +00003906after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb3e7afd2006-03-24 07:16:10 +00003907optimizations. The intended use is to be inserted after optimizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +00003908correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00003909</p>
3910
3911<h5>Arguments:</h5>
3912
3913<p>
3914<tt>id</tt> is a numerical id identifying the marker.
3915</p>
3916
3917<h5>Semantics:</h5>
3918
3919<p>
3920This intrinsic does not modify the behavior of the program. Backends that do not
3921support this intrinisic may ignore it.
3922</p>
3923
3924</div>
3925
Andrew Lenharth51b8d542005-11-11 16:47:30 +00003926<!-- _______________________________________________________________________ -->
3927<div class="doc_subsubsection">
3928 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
3929</div>
3930
3931<div class="doc_text">
3932
3933<h5>Syntax:</h5>
3934<pre>
3935 declare ulong %llvm.readcyclecounter( )
3936</pre>
3937
3938<h5>Overview:</h5>
3939
3940
3941<p>
3942The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
3943counter register (or similar low latency, high accuracy clocks) on those targets
3944that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
3945As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
3946should only be used for small timings.
3947</p>
3948
3949<h5>Semantics:</h5>
3950
3951<p>
3952When directly supported, reading the cycle counter should not modify any memory.
3953Implementations are allowed to either return a application specific value or a
3954system wide value. On backends without support, this is lowered to a constant 0.
3955</p>
3956
3957</div>
3958
Chris Lattner10610642004-02-14 04:08:35 +00003959<!-- ======================================================================= -->
3960<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +00003961 <a name="int_libc">Standard C Library Intrinsics</a>
3962</div>
3963
3964<div class="doc_text">
3965<p>
Chris Lattner10610642004-02-14 04:08:35 +00003966LLVM provides intrinsics for a few important standard C library functions.
3967These intrinsics allow source-language front-ends to pass information about the
3968alignment of the pointer arguments to the code generator, providing opportunity
3969for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003970</p>
3971
3972</div>
3973
3974<!-- _______________________________________________________________________ -->
3975<div class="doc_subsubsection">
3976 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
3977</div>
3978
3979<div class="doc_text">
3980
3981<h5>Syntax:</h5>
3982<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00003983 declare void %llvm.memcpy.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3984 uint &lt;len&gt;, uint &lt;align&gt;)
3985 declare void %llvm.memcpy.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3986 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00003987</pre>
3988
3989<h5>Overview:</h5>
3990
3991<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003992The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00003993location to the destination location.
3994</p>
3995
3996<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00003997Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
3998intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003999</p>
4000
4001<h5>Arguments:</h5>
4002
4003<p>
4004The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00004005the source. The third argument is an integer argument
Chris Lattner33aec9e2004-02-12 17:01:32 +00004006specifying the number of bytes to copy, and the fourth argument is the alignment
4007of the source and destination locations.
4008</p>
4009
Chris Lattner3301ced2004-02-12 21:18:15 +00004010<p>
4011If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00004012the caller guarantees that both the source and destination pointers are aligned
4013to that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00004014</p>
4015
Chris Lattner33aec9e2004-02-12 17:01:32 +00004016<h5>Semantics:</h5>
4017
4018<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004019The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00004020location to the destination location, which are not allowed to overlap. It
4021copies "len" bytes of memory over. If the argument is known to be aligned to
4022some boundary, this can be specified as the fourth argument, otherwise it should
4023be set to 0 or 1.
4024</p>
4025</div>
4026
4027
Chris Lattner0eb51b42004-02-12 18:10:10 +00004028<!-- _______________________________________________________________________ -->
4029<div class="doc_subsubsection">
4030 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
4031</div>
4032
4033<div class="doc_text">
4034
4035<h5>Syntax:</h5>
4036<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00004037 declare void %llvm.memmove.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
4038 uint &lt;len&gt;, uint &lt;align&gt;)
4039 declare void %llvm.memmove.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
4040 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00004041</pre>
4042
4043<h5>Overview:</h5>
4044
4045<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004046The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4047location to the destination location. It is similar to the
4048'<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattner0eb51b42004-02-12 18:10:10 +00004049</p>
4050
4051<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004052Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4053intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner0eb51b42004-02-12 18:10:10 +00004054</p>
4055
4056<h5>Arguments:</h5>
4057
4058<p>
4059The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00004060the source. The third argument is an integer argument
Chris Lattner0eb51b42004-02-12 18:10:10 +00004061specifying the number of bytes to copy, and the fourth argument is the alignment
4062of the source and destination locations.
4063</p>
4064
Chris Lattner3301ced2004-02-12 21:18:15 +00004065<p>
4066If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00004067the caller guarantees that the source and destination pointers are aligned to
4068that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00004069</p>
4070
Chris Lattner0eb51b42004-02-12 18:10:10 +00004071<h5>Semantics:</h5>
4072
4073<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004074The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner0eb51b42004-02-12 18:10:10 +00004075location to the destination location, which may overlap. It
4076copies "len" bytes of memory over. If the argument is known to be aligned to
4077some boundary, this can be specified as the fourth argument, otherwise it should
4078be set to 0 or 1.
4079</p>
4080</div>
4081
Chris Lattner8ff75902004-01-06 05:31:32 +00004082
Chris Lattner10610642004-02-14 04:08:35 +00004083<!-- _______________________________________________________________________ -->
4084<div class="doc_subsubsection">
Chris Lattner5b310c32006-03-03 00:07:20 +00004085 <a name="i_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner10610642004-02-14 04:08:35 +00004086</div>
4087
4088<div class="doc_text">
4089
4090<h5>Syntax:</h5>
4091<pre>
Chris Lattner5b310c32006-03-03 00:07:20 +00004092 declare void %llvm.memset.i32(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
4093 uint &lt;len&gt;, uint &lt;align&gt;)
4094 declare void %llvm.memset.i64(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
4095 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00004096</pre>
4097
4098<h5>Overview:</h5>
4099
4100<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004101The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner10610642004-02-14 04:08:35 +00004102byte value.
4103</p>
4104
4105<p>
4106Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4107does not return a value, and takes an extra alignment argument.
4108</p>
4109
4110<h5>Arguments:</h5>
4111
4112<p>
4113The first argument is a pointer to the destination to fill, the second is the
Chris Lattner5b310c32006-03-03 00:07:20 +00004114byte value to fill it with, the third argument is an integer
Chris Lattner10610642004-02-14 04:08:35 +00004115argument specifying the number of bytes to fill, and the fourth argument is the
4116known alignment of destination location.
4117</p>
4118
4119<p>
4120If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00004121the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner10610642004-02-14 04:08:35 +00004122</p>
4123
4124<h5>Semantics:</h5>
4125
4126<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004127The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4128the
Chris Lattner10610642004-02-14 04:08:35 +00004129destination location. If the argument is known to be aligned to some boundary,
4130this can be specified as the fourth argument, otherwise it should be set to 0 or
41311.
4132</p>
4133</div>
4134
4135
Chris Lattner32006282004-06-11 02:28:03 +00004136<!-- _______________________________________________________________________ -->
4137<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +00004138 <a name="i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a>
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00004139</div>
4140
4141<div class="doc_text">
4142
4143<h5>Syntax:</h5>
4144<pre>
Reid Spencer0b118202006-01-16 21:12:35 +00004145 declare bool %llvm.isunordered.f32(float Val1, float Val2)
4146 declare bool %llvm.isunordered.f64(double Val1, double Val2)
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00004147</pre>
4148
4149<h5>Overview:</h5>
4150
4151<p>
Reid Spencer0b118202006-01-16 21:12:35 +00004152The '<tt>llvm.isunordered</tt>' intrinsics return true if either or both of the
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +00004153specified floating point values is a NAN.
4154</p>
4155
4156<h5>Arguments:</h5>
4157
4158<p>
4159The arguments are floating point numbers of the same type.
4160</p>
4161
4162<h5>Semantics:</h5>
4163
4164<p>
4165If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
4166false.
4167</p>
4168</div>
4169
4170
Chris Lattnera4d74142005-07-21 01:29:16 +00004171<!-- _______________________________________________________________________ -->
4172<div class="doc_subsubsection">
Chris Lattnerec6cb612006-01-16 22:38:59 +00004173 <a name="i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +00004174</div>
4175
4176<div class="doc_text">
4177
4178<h5>Syntax:</h5>
4179<pre>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004180 declare float %llvm.sqrt.f32(float %Val)
4181 declare double %llvm.sqrt.f64(double %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +00004182</pre>
4183
4184<h5>Overview:</h5>
4185
4186<p>
Reid Spencer0b118202006-01-16 21:12:35 +00004187The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Chris Lattnera4d74142005-07-21 01:29:16 +00004188returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
4189<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
4190negative numbers (which allows for better optimization).
4191</p>
4192
4193<h5>Arguments:</h5>
4194
4195<p>
4196The argument and return value are floating point numbers of the same type.
4197</p>
4198
4199<h5>Semantics:</h5>
4200
4201<p>
4202This function returns the sqrt of the specified operand if it is a positive
4203floating point number.
4204</p>
4205</div>
4206
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004207<!-- _______________________________________________________________________ -->
4208<div class="doc_subsubsection">
4209 <a name="i_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
4210</div>
4211
4212<div class="doc_text">
4213
4214<h5>Syntax:</h5>
4215<pre>
4216 declare float %llvm.powi.f32(float %Val, int %power)
4217 declare double %llvm.powi.f64(double %Val, int %power)
4218</pre>
4219
4220<h5>Overview:</h5>
4221
4222<p>
4223The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4224specified (positive or negative) power. The order of evaluation of
4225multiplications is not defined.
4226</p>
4227
4228<h5>Arguments:</h5>
4229
4230<p>
4231The second argument is an integer power, and the first is a value to raise to
4232that power.
4233</p>
4234
4235<h5>Semantics:</h5>
4236
4237<p>
4238This function returns the first value raised to the second power with an
4239unspecified sequence of rounding operations.</p>
4240</div>
4241
4242
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004243<!-- ======================================================================= -->
4244<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +00004245 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004246</div>
4247
4248<div class="doc_text">
4249<p>
Nate Begeman7e36c472006-01-13 23:26:38 +00004250LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004251These allow efficient code generation for some algorithms.
4252</p>
4253
4254</div>
4255
4256<!-- _______________________________________________________________________ -->
4257<div class="doc_subsubsection">
Nate Begeman7e36c472006-01-13 23:26:38 +00004258 <a name="i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
4259</div>
4260
4261<div class="doc_text">
4262
4263<h5>Syntax:</h5>
4264<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00004265 declare ushort %llvm.bswap.i16(ushort &lt;id&gt;)
4266 declare uint %llvm.bswap.i32(uint &lt;id&gt;)
4267 declare ulong %llvm.bswap.i64(ulong &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +00004268</pre>
4269
4270<h5>Overview:</h5>
4271
4272<p>
4273The '<tt>llvm.bwsap</tt>' family of intrinsics is used to byteswap a 16, 32 or
427464 bit quantity. These are useful for performing operations on data that is not
4275in the target's native byte order.
4276</p>
4277
4278<h5>Semantics:</h5>
4279
4280<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +00004281The <tt>llvm.bswap.16</tt> intrinsic returns a ushort value that has the high and low
4282byte of the input ushort swapped. Similarly, the <tt>llvm.bswap.i32</tt> intrinsic
Nate Begeman7e36c472006-01-13 23:26:38 +00004283returns a uint value that has the four bytes of the input uint swapped, so that
4284if the input bytes are numbered 0, 1, 2, 3 then the returned uint will have its
Chris Lattnerec6cb612006-01-16 22:38:59 +00004285bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i64</tt> intrinsic extends this concept
Nate Begeman7e36c472006-01-13 23:26:38 +00004286to 64 bits.
4287</p>
4288
4289</div>
4290
4291<!-- _______________________________________________________________________ -->
4292<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +00004293 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004294</div>
4295
4296<div class="doc_text">
4297
4298<h5>Syntax:</h5>
4299<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00004300 declare ubyte %llvm.ctpop.i8 (ubyte &lt;src&gt;)
4301 declare ushort %llvm.ctpop.i16(ushort &lt;src&gt;)
4302 declare uint %llvm.ctpop.i32(uint &lt;src&gt;)
4303 declare ulong %llvm.ctpop.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004304</pre>
4305
4306<h5>Overview:</h5>
4307
4308<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +00004309The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
4310value.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004311</p>
4312
4313<h5>Arguments:</h5>
4314
4315<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00004316The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00004317unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004318</p>
4319
4320<h5>Semantics:</h5>
4321
4322<p>
4323The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
4324</p>
4325</div>
4326
4327<!-- _______________________________________________________________________ -->
4328<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00004329 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004330</div>
4331
4332<div class="doc_text">
4333
4334<h5>Syntax:</h5>
4335<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00004336 declare ubyte %llvm.ctlz.i8 (ubyte &lt;src&gt;)
4337 declare ushort %llvm.ctlz.i16(ushort &lt;src&gt;)
4338 declare uint %llvm.ctlz.i32(uint &lt;src&gt;)
4339 declare ulong %llvm.ctlz.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004340</pre>
4341
4342<h5>Overview:</h5>
4343
4344<p>
Reid Spencer0b118202006-01-16 21:12:35 +00004345The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
4346leading zeros in a variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004347</p>
4348
4349<h5>Arguments:</h5>
4350
4351<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00004352The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00004353unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004354</p>
4355
4356<h5>Semantics:</h5>
4357
4358<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00004359The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
4360in a variable. If the src == 0 then the result is the size in bits of the type
Chris Lattner99d3c272006-04-21 21:37:40 +00004361of src. For example, <tt>llvm.ctlz(int 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004362</p>
4363</div>
Chris Lattner32006282004-06-11 02:28:03 +00004364
4365
Chris Lattnereff29ab2005-05-15 19:39:26 +00004366
4367<!-- _______________________________________________________________________ -->
4368<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00004369 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +00004370</div>
4371
4372<div class="doc_text">
4373
4374<h5>Syntax:</h5>
4375<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +00004376 declare ubyte %llvm.cttz.i8 (ubyte &lt;src&gt;)
4377 declare ushort %llvm.cttz.i16(ushort &lt;src&gt;)
4378 declare uint %llvm.cttz.i32(uint &lt;src&gt;)
4379 declare ulong %llvm.cttz.i64(ulong &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +00004380</pre>
4381
4382<h5>Overview:</h5>
4383
4384<p>
Reid Spencer0b118202006-01-16 21:12:35 +00004385The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
4386trailing zeros.
Chris Lattnereff29ab2005-05-15 19:39:26 +00004387</p>
4388
4389<h5>Arguments:</h5>
4390
4391<p>
4392The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +00004393unsigned integer type. The return type must match the argument type.
Chris Lattnereff29ab2005-05-15 19:39:26 +00004394</p>
4395
4396<h5>Semantics:</h5>
4397
4398<p>
4399The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
4400in a variable. If the src == 0 then the result is the size in bits of the type
4401of src. For example, <tt>llvm.cttz(2) = 1</tt>.
4402</p>
4403</div>
4404
Chris Lattner8ff75902004-01-06 05:31:32 +00004405<!-- ======================================================================= -->
4406<div class="doc_subsection">
4407 <a name="int_debugger">Debugger Intrinsics</a>
4408</div>
4409
4410<div class="doc_text">
4411<p>
4412The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
4413are described in the <a
4414href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
4415Debugging</a> document.
4416</p>
4417</div>
4418
4419
Chris Lattner00950542001-06-06 20:29:01 +00004420<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00004421<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00004422<address>
4423 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
4424 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
4425 <a href="http://validator.w3.org/check/referer"><img
4426 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
4427
4428 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +00004429 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00004430 Last modified: $Date$
4431</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004432</body>
4433</html>