Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 2a6153684843f76bb098e5c8c361c1e63fea5ccb / . / docs / LangRef.html
blob: c1f37e2b8fee437a6389b1b7829d03cb0dc050b4 [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1<!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 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]6 <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 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]14
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]15<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]17 <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 +0000[diff] [blame]20 <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 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]24 <li><a href="#globalvars">Global Variables</a></li>
25 <li><a href="#functionstructure">Function Structure</a></li>
26 </ol>
27 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]28 <li><a href="#typesystem">Type System</a>
29 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]30 <li><a href="#t_primitive">Primitive Types</a>
31 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]32 <li><a href="#t_classifications">Type Classifications</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]33 </ol>
34 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]35 <li><a href="#t_derived">Derived Types</a>
36 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]37 <li><a href="#t_array">Array Type</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]38 <li><a href="#t_function">Function Type</a></li>
39 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]40 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]41 <li><a href="#t_packed">Packed Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]42 </ol>
43 </li>
44 </ol>
45 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]46 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]47 <ol>
48 <li><a href="#simpleconstants">Simple Constants</a>
49 <li><a href="#aggregateconstants">Aggregate Constants</a>
50 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
51 <li><a href="#undefvalues">Undefined Values</a>
52 <li><a href="#constantexprs">Constant Expressions</a>
53 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]54 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]55 <li><a href="#instref">Instruction Reference</a>
56 <ol>
57 <li><a href="#terminators">Terminator Instructions</a>
58 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]59 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
60 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]61 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
62 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]63 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]64 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]65 </ol>
66 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]67 <li><a href="#binaryops">Binary Operations</a>
68 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]69 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
70 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
71 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
72 <li><a href="#i_div">'<tt>div</tt>' Instruction</a></li>
73 <li><a href="#i_rem">'<tt>rem</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]74 <li><a href="#i_setcc">'<tt>set<i>cc</i></tt>' Instructions</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]75 </ol>
76 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]77 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
78 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]79 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]80 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]81 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
82 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
83 <li><a href="#i_shr">'<tt>shr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]84 </ol>
85 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]86 <li><a href="#memoryops">Memory Access Operations</a>
87 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]88 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
89 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
90 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
91 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
92 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
93 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
94 </ol>
95 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]96 <li><a href="#otherops">Other Operations</a>
97 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]98 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]99 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]100 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]101 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]102 <li><a href="#i_vanext">'<tt>vanext</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]103 <li><a href="#i_vaarg">'<tt>vaarg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]104 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]105 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]106 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]107 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]108 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]109 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]110 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
111 <ol>
112 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
113 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
114 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
115 </ol>
116 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]117 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
118 <ol>
119 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
120 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
121 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
122 </ol>
123 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]124 <li><a href="#int_codegen">Code Generator Intrinsics</a>
125 <ol>
126 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
127 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]128 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]129 </ol>
130 </li>
131 <li><a href="#int_os">Operating System Intrinsics</a>
132 <ol>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]133 <li><a href="#i_readport">'<tt>llvm.readport</tt>' Intrinsic</a></li>
134 <li><a href="#i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a></li>
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]135 <li><a href="#i_readio">'<tt>llvm.readio</tt>' Intrinsic</a></li>
136 <li><a href="#i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]137 </ol>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]138 <li><a href="#int_libc">Standard C Library Intrinsics</a>
139 <ol>
140 <li><a href="#i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a></li>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]141 <li><a href="#i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]142 <li><a href="#i_memset">'<tt>llvm.memset</tt>' Intrinsic</a></li>
Alkis Evlogimenos96853722004-06-12 19:19:14 +0000[diff] [blame]143 <li><a href="#i_isunordered">'<tt>llvm.isunordered</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]144 </ol>
145 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]146 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]147 </ol>
148 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]149</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]150
151<div class="doc_author">
152 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
153 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]154</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]155
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]156<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]157<div class="doc_section"> <a name="abstract">Abstract </a></div>
158<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]159
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]160<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]161<p>This document is a reference manual for the LLVM assembly language.
162LLVM is an SSA based representation that provides type safety,
163low-level operations, flexibility, and the capability of representing
164'all' high-level languages cleanly. It is the common code
165representation used throughout all phases of the LLVM compilation
166strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]167</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]168
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]169<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]170<div class="doc_section"> <a name="introduction">Introduction</a> </div>
171<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]172
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]173<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]174
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]175<p>The LLVM code representation is designed to be used in three
176different forms: as an in-memory compiler IR, as an on-disk bytecode
177representation (suitable for fast loading by a Just-In-Time compiler),
178and as a human readable assembly language representation. This allows
179LLVM to provide a powerful intermediate representation for efficient
180compiler transformations and analysis, while providing a natural means
181to debug and visualize the transformations. The three different forms
182of LLVM are all equivalent. This document describes the human readable
183representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]184
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]185<p>The LLVM representation aims to be a light-weight and low-level
186while being expressive, typed, and extensible at the same time. It
187aims to be a "universal IR" of sorts, by being at a low enough level
188that high-level ideas may be cleanly mapped to it (similar to how
189microprocessors are "universal IR's", allowing many source languages to
190be mapped to them). By providing type information, LLVM can be used as
191the target of optimizations: for example, through pointer analysis, it
192can be proven that a C automatic variable is never accessed outside of
193the current function... allowing it to be promoted to a simple SSA
194value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]195
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]196</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]197
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]198<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]199<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]200
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]201<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]202
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]203<p>It is important to note that this document describes 'well formed'
204LLVM assembly language. There is a difference between what the parser
205accepts and what is considered 'well formed'. For example, the
206following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]207
208<pre>
209 %x = <a href="#i_add">add</a> int 1, %x
210</pre>
211
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]212<p>...because the definition of <tt>%x</tt> does not dominate all of
213its uses. The LLVM infrastructure provides a verification pass that may
214be used to verify that an LLVM module is well formed. This pass is
215automatically run by the parser after parsing input assembly, and by
216the optimizer before it outputs bytecode. The violations pointed out
217by the verifier pass indicate bugs in transformation passes or input to
218the parser.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]219
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]220<!-- Describe the typesetting conventions here. --> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]221
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]222<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]223<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]224<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]225
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]226<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]227
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]228<p>LLVM uses three different forms of identifiers, for different
229purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]230
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]231<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]232 <li>Named values are represented as a string of characters with a '%' prefix.
233 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
234 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
235 Identifiers which require other characters in their names can be surrounded
236 with quotes. In this way, anything except a <tt>"</tt> character can be used
237 in a name.</li>
238
239 <li>Unnamed values are represented as an unsigned numeric value with a '%'
240 prefix. For example, %12, %2, %44.</li>
241
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]242 <li>Constants, which are described in a <a href="#constants">section about
243 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]244</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]245
246<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
247don't need to worry about name clashes with reserved words, and the set of
248reserved words may be expanded in the future without penalty. Additionally,
249unnamed identifiers allow a compiler to quickly come up with a temporary
250variable without having to avoid symbol table conflicts.</p>
251
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]252<p>Reserved words in LLVM are very similar to reserved words in other
253languages. There are keywords for different opcodes ('<tt><a
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]254href="#i_add">add</a></tt>', '<tt><a href="#i_cast">cast</a></tt>', '<tt><a
255href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
256href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
257and others. These reserved words cannot conflict with variable names, because
258none of them start with a '%' character.</p>
259
260<p>Here is an example of LLVM code to multiply the integer variable
261'<tt>%X</tt>' by 8:</p>
262
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]263<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]264
265<pre>
266 %result = <a href="#i_mul">mul</a> uint %X, 8
267</pre>
268
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]269<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]270
271<pre>
272 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
273</pre>
274
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]275<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]276
277<pre>
278 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
279 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
280 %result = <a href="#i_add">add</a> uint %1, %1
281</pre>
282
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]283<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
284important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]285
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]286<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]287
288 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
289 line.</li>
290
291 <li>Unnamed temporaries are created when the result of a computation is not
292 assigned to a named value.</li>
293
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]294 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]295
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]296</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]297
298<p>...and it also show a convention that we follow in this document. When
299demonstrating instructions, we will follow an instruction with a comment that
300defines the type and name of value produced. Comments are shown in italic
301text.</p>
302
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]303</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]304
305<!-- *********************************************************************** -->
306<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
307<!-- *********************************************************************** -->
308
309<!-- ======================================================================= -->
310<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
311</div>
312
313<div class="doc_text">
314
315<p>LLVM programs are composed of "Module"s, each of which is a
316translation unit of the input programs. Each module consists of
317functions, global variables, and symbol table entries. Modules may be
318combined together with the LLVM linker, which merges function (and
319global variable) definitions, resolves forward declarations, and merges
320symbol table entries. Here is an example of the "hello world" module:</p>
321
322<pre><i>; Declare the string constant as a global constant...</i>
323<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
324 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
325
326<i>; External declaration of the puts function</i>
327<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
328
329<i>; Definition of main function</i>
330int %main() { <i>; int()* </i>
331 <i>; Convert [13x sbyte]* to sbyte *...</i>
332 %cast210 = <a
333 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
334
335 <i>; Call puts function to write out the string to stdout...</i>
336 <a
337 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
338 <a
339 href="#i_ret">ret</a> int 0<br>}<br></pre>
340
341<p>This example is made up of a <a href="#globalvars">global variable</a>
342named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
343function, and a <a href="#functionstructure">function definition</a>
344for "<tt>main</tt>".</p>
345
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]346<p>In general, a module is made up of a list of global values,
347where both functions and global variables are global values. Global values are
348represented by a pointer to a memory location (in this case, a pointer to an
349array of char, and a pointer to a function), and have one of the following <a
350href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]351
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]352</div>
353
354<!-- ======================================================================= -->
355<div class="doc_subsection">
356 <a name="linkage">Linkage Types</a>
357</div>
358
359<div class="doc_text">
360
361<p>
362All Global Variables and Functions have one of the following types of linkage:
363</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]364
365<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]366
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]367 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]368
369 <dd>Global values with internal linkage are only directly accessible by
370 objects in the current module. In particular, linking code into a module with
371 an internal global value may cause the internal to be renamed as necessary to
372 avoid collisions. Because the symbol is internal to the module, all
373 references can be updated. This corresponds to the notion of the
374 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]375 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]376
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]377 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]378
379 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
380 the twist that linking together two modules defining the same
381 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
382 is typically used to implement inline functions. Unreferenced
383 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]384 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]385
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]386 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]387
388 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
389 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
390 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]391 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]392
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]393 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]394
395 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
396 pointer to array type. When two global variables with appending linkage are
397 linked together, the two global arrays are appended together. This is the
398 LLVM, typesafe, equivalent of having the system linker append together
399 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]400 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]401
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]402 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]403
404 <dd>If none of the above identifiers are used, the global is externally
405 visible, meaning that it participates in linkage and can be used to resolve
406 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]407 </dd>
408</dl>
409
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]410<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
411variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
412variable and was linked with this one, one of the two would be renamed,
413preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
414external (i.e., lacking any linkage declarations), they are accessible
415outside of the current module. It is illegal for a function <i>declaration</i>
416to have any linkage type other than "externally visible".</a></p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]417
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]418</div>
419
420<!-- ======================================================================= -->
421<div class="doc_subsection">
422 <a name="globalvars">Global Variables</a>
423</div>
424
425<div class="doc_text">
426
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]427<p>Global variables define regions of memory allocated at compilation time
428instead of run-time. Global variables may optionally be initialized. A
429variable may be defined as a global "constant", which indicates that the
430contents of the variable will <b>never</b> be modified (enabling better
431optimization, allowing the global data to be placed in the read-only section of
432an executable, etc). Note that variables that need runtime initialization
433cannot be marked "constant", as there is a store to the variable.</p>
434
435<p>
436LLVM explicitly allows <em>declarations</em> of global variables to be marked
437constant, even if the final definition of the global is not. This capability
438can be used to enable slightly better optimization of the program, but requires
439the language definition to guarantee that optimizations based on the
440'constantness' are valid for the translation units that do not include the
441definition.
442</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]443
444<p>As SSA values, global variables define pointer values that are in
445scope (i.e. they dominate) all basic blocks in the program. Global
446variables always define a pointer to their "content" type because they
447describe a region of memory, and all memory objects in LLVM are
448accessed through pointers.</p>
449
450</div>
451
452
453<!-- ======================================================================= -->
454<div class="doc_subsection">
455 <a name="functionstructure">Functions</a>
456</div>
457
458<div class="doc_text">
459
460<p>LLVM function definitions are composed of a (possibly empty) argument list,
461an opening curly brace, a list of basic blocks, and a closing curly brace. LLVM
462function declarations are defined with the "<tt>declare</tt>" keyword, a
463function name, and a function signature.</p>
464
465<p>A function definition contains a list of basic blocks, forming the CFG for
466the function. Each basic block may optionally start with a label (giving the
467basic block a symbol table entry), contains a list of instructions, and ends
468with a <a href="#terminators">terminator</a> instruction (such as a branch or
469function return).</p>
470
471<p>The first basic block in program is special in two ways: it is immediately
472executed on entrance to the function, and it is not allowed to have predecessor
473basic blocks (i.e. there can not be any branches to the entry block of a
474function). Because the block can have no predecessors, it also cannot have any
475<a href="#i_phi">PHI nodes</a>.</p>
476
477<p>LLVM functions are identified by their name and type signature. Hence, two
478functions with the same name but different parameter lists or return values are
479considered different functions, and LLVM will resolves references to each
480appropriately.</p>
481
482</div>
483
484
485
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]486<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]487<div class="doc_section"> <a name="typesystem">Type System</a> </div>
488<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]489
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]490<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]491
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]492<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]493intermediate representation. Being typed enables a number of
494optimizations to be performed on the IR directly, without having to do
495extra analyses on the side before the transformation. A strong type
496system makes it easier to read the generated code and enables novel
497analyses and transformations that are not feasible to perform on normal
498three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]499
500</div>
501
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]502<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]503<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]504<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]505<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]506system. The current set of primitive types are as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]507
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]508<table class="layout">
509 <tr class="layout">
510 <td class="left">
511 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]512 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]513 <tr><th>Type</th><th>Description</th></tr>
514 <tr><td><tt>void</tt></td><td>No value</td></tr>
515 <tr><td><tt>ubyte</tt></td><td>Unsigned 8 bit value</td></tr>
516 <tr><td><tt>ushort</tt></td><td>Unsigned 16 bit value</td></tr>
517 <tr><td><tt>uint</tt></td><td>Unsigned 32 bit value</td></tr>
518 <tr><td><tt>ulong</tt></td><td>Unsigned 64 bit value</td></tr>
519 <tr><td><tt>float</tt></td><td>32 bit floating point value</td></tr>
520 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]521 </tbody>
522 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]523 </td>
524 <td class="right">
525 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]526 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]527 <tr><th>Type</th><th>Description</th></tr>
528 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
529 <tr><td><tt>sbyte</tt></td><td>Signed 8 bit value</td></tr>
530 <tr><td><tt>short</tt></td><td>Signed 16 bit value</td></tr>
531 <tr><td><tt>int</tt></td><td>Signed 32 bit value</td></tr>
532 <tr><td><tt>long</tt></td><td>Signed 64 bit value</td></tr>
533 <tr><td><tt>double</tt></td><td>64 bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]534 </tbody>
535 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]536 </td>
537 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]538</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]539</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]540
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]541<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]542<div class="doc_subsubsection"> <a name="t_classifications">Type
543Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]544<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]545<p>These different primitive types fall into a few useful
546classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]547
548<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]549 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]550 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]551 <tr>
552 <td><a name="t_signed">signed</a></td>
553 <td><tt>sbyte, short, int, long, float, double</tt></td>
554 </tr>
555 <tr>
556 <td><a name="t_unsigned">unsigned</a></td>
557 <td><tt>ubyte, ushort, uint, ulong</tt></td>
558 </tr>
559 <tr>
560 <td><a name="t_integer">integer</a></td>
561 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
562 </tr>
563 <tr>
564 <td><a name="t_integral">integral</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]565 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
566 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]567 </tr>
568 <tr>
569 <td><a name="t_floating">floating point</a></td>
570 <td><tt>float, double</tt></td>
571 </tr>
572 <tr>
573 <td><a name="t_firstclass">first class</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]574 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
575 float, double, <a href="#t_pointer">pointer</a>,
576 <a href="#t_packed">packed</a></tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]577 </tr>
578 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]579</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]580
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]581<p>The <a href="#t_firstclass">first class</a> types are perhaps the
582most important. Values of these types are the only ones which can be
583produced by instructions, passed as arguments, or used as operands to
584instructions. This means that all structures and arrays must be
585manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]586</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]587
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]588<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]589<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]590
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]591<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]592
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]593<p>The real power in LLVM comes from the derived types in the system.
594This is what allows a programmer to represent arrays, functions,
595pointers, and other useful types. Note that these derived types may be
596recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]597
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]598</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]599
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]600<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]601<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]602
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]603<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]604
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]605<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]606
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]607<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]608sequentially in memory. The array type requires a size (number of
609elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]610
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]611<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]612
613<pre>
614 [&lt;# elements&gt; x &lt;elementtype&gt;]
615</pre>
616
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]617<p>The number of elements is a constant integer value, elementtype may
618be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]619
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]620<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]621<table class="layout">
622 <tr class="layout">
623 <td class="left">
624 <tt>[40 x int ]</tt><br/>
625 <tt>[41 x int ]</tt><br/>
626 <tt>[40 x uint]</tt><br/>
627 </td>
628 <td class="left">
629 Array of 40 integer values.<br/>
630 Array of 41 integer values.<br/>
631 Array of 40 unsigned integer values.<br/>
632 </td>
633 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]634</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]635<p>Here are some examples of multidimensional arrays:</p>
636<table class="layout">
637 <tr class="layout">
638 <td class="left">
639 <tt>[3 x [4 x int]]</tt><br/>
640 <tt>[12 x [10 x float]]</tt><br/>
641 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
642 </td>
643 <td class="left">
644 3x4 array integer values.<br/>
645 12x10 array of single precision floating point values.<br/>
646 2x3x4 array of unsigned integer values.<br/>
647 </td>
648 </tr>
649</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]650</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]651
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]652<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]653<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]654<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]655<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]656<p>The function type can be thought of as a function signature. It
657consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]658Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]659(which are structures of pointers to functions), for indirect function
660calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]661<p>
662The return type of a function type cannot be an aggregate type.
663</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]664<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]665<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]666<p>Where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
667specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]668which indicates that the function takes a variable number of arguments.
669Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]670 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]671<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]672<table class="layout">
673 <tr class="layout">
674 <td class="left">
675 <tt>int (int)</tt> <br/>
676 <tt>float (int, int *) *</tt><br/>
677 <tt>int (sbyte *, ...)</tt><br/>
678 </td>
679 <td class="left">
680 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
681 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]682 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]683 returning <tt>float</tt>.<br/>
684 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
685 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
686 the signature for <tt>printf</tt> in LLVM.<br/>
687 </td>
688 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]689</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]690
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]691</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]692<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]693<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]694<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]695<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]696<p>The structure type is used to represent a collection of data members
697together in memory. The packing of the field types is defined to match
698the ABI of the underlying processor. The elements of a structure may
699be any type that has a size.</p>
700<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
701and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
702field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
703instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]704<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]705<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]706<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]707<table class="layout">
708 <tr class="layout">
709 <td class="left">
710 <tt>{ int, int, int }</tt><br/>
711 <tt>{ float, int (int) * }</tt><br/>
712 </td>
713 <td class="left">
714 a triple of three <tt>int</tt> values<br/>
715 A pair, where the first element is a <tt>float</tt> and the second element
716 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
717 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
718 </td>
719 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]720</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]721</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]722
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]723<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]724<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]725<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]726<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]727<p>As in many languages, the pointer type represents a pointer or
728reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]729<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]730<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]731<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]732<table class="layout">
733 <tr class="layout">
734 <td class="left">
735 <tt>[4x int]*</tt><br/>
736 <tt>int (int *) *</tt><br/>
737 </td>
738 <td class="left">
739 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
740 four <tt>int</tt> values<br/>
741 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattnera977c482005-02-19 02:22:14 +0000[diff] [blame]742 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]743 <tt>int</tt>.<br/>
744 </td>
745 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]746</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]747</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]748
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]749<!-- _______________________________________________________________________ -->
750<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]751<div class="doc_text">
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]752<h5>Overview:</h5>
753<p>A packed type is a simple derived type that represents a vector
754of elements. Packed types are used when multiple primitive data
755are operated in parallel using a single instruction (SIMD).
756A packed type requires a size (number of
757elements) and an underlying primitive data type. Packed types are
758considered <a href="#t_firstclass">first class</a>.</p>
759<h5>Syntax:</h5>
760<pre> &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;<br></pre>
761<p>The number of elements is a constant integer value, elementtype may
762be any integral or floating point type.</p>
763<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]764<table class="layout">
765 <tr class="layout">
766 <td class="left">
767 <tt>&lt;4 x int&gt;</tt><br/>
768 <tt>&lt;8 x float&gt;</tt><br/>
769 <tt>&lt;2 x uint&gt;</tt><br/>
770 </td>
771 <td class="left">
772 Packed vector of 4 integer values.<br/>
773 Packed vector of 8 floating-point values.<br/>
774 Packed vector of 2 unsigned integer values.<br/>
775 </td>
776 </tr>
777</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]778</div>
779
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]780<!-- *********************************************************************** -->
781<div class="doc_section"> <a name="constants">Constants</a> </div>
782<!-- *********************************************************************** -->
783
784<div class="doc_text">
785
786<p>LLVM has several different basic types of constants. This section describes
787them all and their syntax.</p>
788
789</div>
790
791<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]792<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]793
794<div class="doc_text">
795
796<dl>
797 <dt><b>Boolean constants</b></dt>
798
799 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
800 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
801 </dd>
802
803 <dt><b>Integer constants</b></dt>
804
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]805 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]806 href="#t_integer">integer</a> type. Negative numbers may be used with signed
807 integer types.
808 </dd>
809
810 <dt><b>Floating point constants</b></dt>
811
812 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
813 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]814 notation. Floating point constants have an optional hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]815 notation (see below). Floating point constants must have a <a
816 href="#t_floating">floating point</a> type. </dd>
817
818 <dt><b>Null pointer constants</b></dt>
819
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]820 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]821 and must be of <a href="#t_pointer">pointer type</a>.</dd>
822
823</dl>
824
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]825<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]826of floating point constants. For example, the form '<tt>double
8270x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
8284.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]829(and the only time that they are generated by the disassembler) is when a
830floating point constant must be emitted but it cannot be represented as a
831decimal floating point number. For example, NaN's, infinities, and other
832special values are represented in their IEEE hexadecimal format so that
833assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]834
835</div>
836
837<!-- ======================================================================= -->
838<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
839</div>
840
841<div class="doc_text">
842
843<dl>
844 <dt><b>Structure constants</b></dt>
845
846 <dd>Structure constants are represented with notation similar to structure
847 type definitions (a comma separated list of elements, surrounded by braces
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]848 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0 }</tt>". Structure
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]849 constants must have <a href="#t_struct">structure type</a>, and the number and
850 types of elements must match those specified by the type.
851 </dd>
852
853 <dt><b>Array constants</b></dt>
854
855 <dd>Array constants are represented with notation similar to array type
856 definitions (a comma separated list of elements, surrounded by square brackets
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]857 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]858 constants must have <a href="#t_array">array type</a>, and the number and
859 types of elements must match those specified by the type.
860 </dd>
861
862 <dt><b>Packed constants</b></dt>
863
864 <dd>Packed constants are represented with notation similar to packed type
865 definitions (a comma separated list of elements, surrounded by
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]866 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; int 42,
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]867 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
868 href="#t_packed">packed type</a>, and the number and types of elements must
869 match those specified by the type.
870 </dd>
871
872 <dt><b>Zero initialization</b></dt>
873
874 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
875 value to zero of <em>any</em> type, including scalar and aggregate types.
876 This is often used to avoid having to print large zero initializers (e.g. for
877 large arrays), and is always exactly equivalent to using explicit zero
878 initializers.
879 </dd>
880</dl>
881
882</div>
883
884<!-- ======================================================================= -->
885<div class="doc_subsection">
886 <a name="globalconstants">Global Variable and Function Addresses</a>
887</div>
888
889<div class="doc_text">
890
891<p>The addresses of <a href="#globalvars">global variables</a> and <a
892href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]893constants. These constants are explicitly referenced when the <a
894href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]895href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
896file:</p>
897
898<pre>
899 %X = global int 17
900 %Y = global int 42
901 %Z = global [2 x int*] [ int* %X, int* %Y ]
902</pre>
903
904</div>
905
906<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]907<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]908<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]909 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
910 no specific value. Undefined values may be of any type, and be used anywhere
911 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]912
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]913 <p>Undefined values indicate to the compiler that the program is well defined
914 no matter what value is used, giving the compiler more freedom to optimize.
915 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]916</div>
917
918<!-- ======================================================================= -->
919<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
920</div>
921
922<div class="doc_text">
923
924<p>Constant expressions are used to allow expressions involving other constants
925to be used as constants. Constant expressions may be of any <a
926href="#t_firstclass">first class</a> type, and may involve any LLVM operation
927that does not have side effects (e.g. load and call are not supported). The
928following is the syntax for constant expressions:</p>
929
930<dl>
931 <dt><b><tt>cast ( CST to TYPE )</tt></b></dt>
932
933 <dd>Cast a constant to another type.</dd>
934
935 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
936
937 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
938 constants. As with the <a href="#i_getelementptr">getelementptr</a>
939 instruction, the index list may have zero or more indexes, which are required
940 to make sense for the type of "CSTPTR".</dd>
941
942 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
943
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]944 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
945 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]946 binary</a> operations. The constraints on operands are the same as those for
947 the corresponding instruction (e.g. no bitwise operations on floating point
948 are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]949</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]950</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]951
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]952<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]953<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
954<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]955
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]956<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]957
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]958<p>The LLVM instruction set consists of several different
959classifications of instructions: <a href="#terminators">terminator
960instructions</a>, <a href="#binaryops">binary instructions</a>, <a
961 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
962instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]963
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]964</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]965
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]966<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]967<div class="doc_subsection"> <a name="terminators">Terminator
968Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]969
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]970<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]971
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]972<p>As mentioned <a href="#functionstructure">previously</a>, every
973basic block in a program ends with a "Terminator" instruction, which
974indicates which block should be executed after the current block is
975finished. These terminator instructions typically yield a '<tt>void</tt>'
976value: they produce control flow, not values (the one exception being
977the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]978<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]979 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
980instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]981the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
982 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
983 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]984
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]985</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]986
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]987<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]988<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
989Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]990<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]991<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]992<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 +0000[diff] [blame]993 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]994</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]995<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]996<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
997value) from a function, back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]998<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]999returns a value and then causes control flow, and one that just causes
1000control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1001<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1002<p>The '<tt>ret</tt>' instruction may return any '<a
1003 href="#t_firstclass">first class</a>' type. Notice that a function is
1004not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1005instruction inside of the function that returns a value that does not
1006match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1007<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1008<p>When the '<tt>ret</tt>' instruction is executed, control flow
1009returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +0000[diff] [blame]1010 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1011the instruction after the call. If the caller was an "<a
1012 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
1013at the beginning "normal" of the destination block. If the instruction
1014returns a value, that value shall set the call or invoke instruction's
1015return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1016<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1017<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1018 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1019</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1020</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1021<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1022<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1023<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1024<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1025<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 +0000[diff] [blame]1026</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1027<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1028<p>The '<tt>br</tt>' instruction is used to cause control flow to
1029transfer to a different basic block in the current function. There are
1030two forms of this instruction, corresponding to a conditional branch
1031and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1032<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1033<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1034single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1035unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1036value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1037<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1038<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1039argument is evaluated. If the value is <tt>true</tt>, control flows
1040to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1041control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1042<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1043<pre>Test:<br> %cond = <a href="#i_setcc">seteq</a> int %a, %b<br> br bool %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
1044 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 +0000[diff] [blame]1045</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1046<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1047<div class="doc_subsubsection">
1048 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1049</div>
1050
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1051<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1052<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1053
1054<pre>
1055 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1056</pre>
1057
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1058<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1059
1060<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1061several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1062instruction, allowing a branch to occur to one of many possible
1063destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1064
1065
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1066<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1067
1068<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1069comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1070an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1071table is not allowed to contain duplicate constant entries.</p>
1072
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1073<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1074
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1075<p>The <tt>switch</tt> instruction specifies a table of values and
1076destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1077table is searched for the given value. If the value is found, control flow is
1078transfered to the corresponding destination; otherwise, control flow is
1079transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1080
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1081<h5>Implementation:</h5>
1082
1083<p>Depending on properties of the target machine and the particular
1084<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1085ways. For example, it could be generated as a series of chained conditional
1086branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1087
1088<h5>Example:</h5>
1089
1090<pre>
1091 <i>; Emulate a conditional br instruction</i>
1092 %Val = <a href="#i_cast">cast</a> bool %value to int
1093 switch int %Val, label %truedest [int 0, label %falsedest ]
1094
1095 <i>; Emulate an unconditional br instruction</i>
1096 switch uint 0, label %dest [ ]
1097
1098 <i>; Implement a jump table:</i>
1099 switch uint %val, label %otherwise [ uint 0, label %onzero
1100 uint 1, label %onone
1101 uint 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1102</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1103</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1104<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1105<div class="doc_subsubsection"> <a name="i_invoke">'<tt>invoke</tt>'
1106Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1107<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1108<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1109<pre> &lt;result&gt; = invoke &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)<br> to label &lt;normal label&gt; except label &lt;exception label&gt;<br></pre>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1110<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1111<p>The '<tt>invoke</tt>' instruction causes control to transfer to a
1112specified function, with the possibility of control flow transfer to
1113either the '<tt>normal</tt>' <tt>label</tt> label or the '<tt>exception</tt>'<tt>label</tt>.
1114If the callee function returns with the "<tt><a href="#i_ret">ret</a></tt>"
1115instruction, control flow will return to the "normal" label. If the
1116callee (or any indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
1117instruction, control is interrupted, and continued at the dynamically
1118nearest "except" label.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1119<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1120<p>This instruction requires several arguments:</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1121<ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1122 <li>'<tt>ptr to function ty</tt>': shall be the signature of the
1123pointer to function value being invoked. In most cases, this is a
1124direct function invocation, but indirect <tt>invoke</tt>s are just as
1125possible, branching off an arbitrary pointer to function value. </li>
1126 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer
1127to a function to be invoked. </li>
1128 <li>'<tt>function args</tt>': argument list whose types match the
1129function signature argument types. If the function signature indicates
1130the function accepts a variable number of arguments, the extra
1131arguments can be specified. </li>
1132 <li>'<tt>normal label</tt>': the label reached when the called
1133function executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1134 <li>'<tt>exception label</tt>': the label reached when a callee
1135returns with the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1136</ol>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1137<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1138<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1139 href="#i_call">call</a></tt>' instruction in most regards. The
1140primary difference is that it establishes an association with a label,
1141which is used by the runtime library to unwind the stack.</p>
1142<p>This instruction is used in languages with destructors to ensure
1143that proper cleanup is performed in the case of either a <tt>longjmp</tt>
1144or a thrown exception. Additionally, this is important for
1145implementation of '<tt>catch</tt>' clauses in high-level languages that
1146support them.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1147<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1148<pre> %retval = invoke int %Test(int 15)<br> to label %Continue<br> except label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1149</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1150</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1151
1152
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1153<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1154
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1155<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1156Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1157
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1158<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1159
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1160<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1161<pre>
1162 unwind
1163</pre>
1164
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1165<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1166
1167<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1168at the first callee in the dynamic call stack which used an <a
1169href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1170primarily used to implement exception handling.</p>
1171
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1172<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1173
1174<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1175immediately halt. The dynamic call stack is then searched for the first <a
1176href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1177execution continues at the "exceptional" destination block specified by the
1178<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1179dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1180</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1181
1182<!-- _______________________________________________________________________ -->
1183
1184<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1185Instruction</a> </div>
1186
1187<div class="doc_text">
1188
1189<h5>Syntax:</h5>
1190<pre>
1191 unreachable
1192</pre>
1193
1194<h5>Overview:</h5>
1195
1196<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1197instruction is used to inform the optimizer that a particular portion of the
1198code is not reachable. This can be used to indicate that the code after a
1199no-return function cannot be reached, and other facts.</p>
1200
1201<h5>Semantics:</h5>
1202
1203<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1204</div>
1205
1206
1207
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1208<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1209<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1210<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1211<p>Binary operators are used to do most of the computation in a
1212program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1213produce a single value. The operands might represent
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1214multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1215The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1216necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1217<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1218</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1219<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1220<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1221Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1222<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1223<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1224<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 +0000[diff] [blame]1225</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1226<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1227<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1228<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1229<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1230 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1231 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1232Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1233<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1234<p>The value produced is the integer or floating point sum of the two
1235operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1236<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1237<pre> &lt;result&gt; = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1238</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1239</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1240<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1241<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1242Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1243<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1244<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1245<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 +0000[diff] [blame]1246</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1247<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1248<p>The '<tt>sub</tt>' instruction returns the difference of its two
1249operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1250<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1251instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1252<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1253<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1254 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1255values.
1256This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1257Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1258<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1259<p>The value produced is the integer or floating point difference of
1260the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1261<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1262<pre> &lt;result&gt; = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1263 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1264</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1265</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1266<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1267<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1268Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1269<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1270<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1271<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 +0000[diff] [blame]1272</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1273<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1274<p>The '<tt>mul</tt>' instruction returns the product of its two
1275operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1276<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1277<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1278 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1279values.
1280This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1281Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1282<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1283<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1284two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1285<p>There is no signed vs unsigned multiplication. The appropriate
1286action is taken based on the type of the operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1287<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1288<pre> &lt;result&gt; = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1289</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1290</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1291<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1292<div class="doc_subsubsection"> <a name="i_div">'<tt>div</tt>'
1293Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1294<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1295<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1296<pre> &lt;result&gt; = div &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1297</pre>
1298<h5>Overview:</h5>
1299<p>The '<tt>div</tt>' instruction returns the quotient of its two
1300operands.</p>
1301<h5>Arguments:</h5>
1302<p>The two arguments to the '<tt>div</tt>' instruction must be either <a
1303 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1304values.
1305This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1306Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1307<h5>Semantics:</h5>
1308<p>The value produced is the integer or floating point quotient of the
1309two operands.</p>
1310<h5>Example:</h5>
1311<pre> &lt;result&gt; = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1312</pre>
1313</div>
1314<!-- _______________________________________________________________________ -->
1315<div class="doc_subsubsection"> <a name="i_rem">'<tt>rem</tt>'
1316Instruction</a> </div>
1317<div class="doc_text">
1318<h5>Syntax:</h5>
1319<pre> &lt;result&gt; = rem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1320</pre>
1321<h5>Overview:</h5>
1322<p>The '<tt>rem</tt>' instruction returns the remainder from the
1323division of its two operands.</p>
1324<h5>Arguments:</h5>
1325<p>The two arguments to the '<tt>rem</tt>' instruction must be either <a
1326 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1327values.
1328This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1329Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1330<h5>Semantics:</h5>
1331<p>This returns the <i>remainder</i> of a division (where the result
1332has the same sign as the divisor), not the <i>modulus</i> (where the
1333result has the same sign as the dividend) of a value. For more
1334information about the difference, see: <a
1335 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1336Math Forum</a>.</p>
1337<h5>Example:</h5>
1338<pre> &lt;result&gt; = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1339</pre>
1340</div>
1341<!-- _______________________________________________________________________ -->
1342<div class="doc_subsubsection"> <a name="i_setcc">'<tt>set<i>cc</i></tt>'
1343Instructions</a> </div>
1344<div class="doc_text">
1345<h5>Syntax:</h5>
1346<pre> &lt;result&gt; = seteq &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1347 &lt;result&gt; = setne &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1348 &lt;result&gt; = setlt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1349 &lt;result&gt; = setgt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1350 &lt;result&gt; = setle &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1351 &lt;result&gt; = setge &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1352</pre>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1353<h5>Overview:</h5>
1354<p>The '<tt>set<i>cc</i></tt>' family of instructions returns a boolean
1355value based on a comparison of their two operands.</p>
1356<h5>Arguments:</h5>
1357<p>The two arguments to the '<tt>set<i>cc</i></tt>' instructions must
1358be of <a href="#t_firstclass">first class</a> type (it is not possible
1359to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1360or '<tt>void</tt>' values, etc...). Both arguments must have identical
1361types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1362<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1363<p>The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1364value if both operands are equal.<br>
1365The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1366value if both operands are unequal.<br>
1367The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1368value if the first operand is less than the second operand.<br>
1369The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1370value if the first operand is greater than the second operand.<br>
1371The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1372value if the first operand is less than or equal to the second operand.<br>
1373The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1374value if the first operand is greater than or equal to the second
1375operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1376<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1377<pre> &lt;result&gt; = seteq int 4, 5 <i>; yields {bool}:result = false</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1378 &lt;result&gt; = setne float 4, 5 <i>; yields {bool}:result = true</i>
1379 &lt;result&gt; = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1380 &lt;result&gt; = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1381 &lt;result&gt; = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1382 &lt;result&gt; = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1383</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1384</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1385<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1386<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1387Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1388<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1389<p>Bitwise binary operators are used to do various forms of
1390bit-twiddling in a program. They are generally very efficient
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1391instructions and can commonly be strength reduced from other
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1392instructions. They require two operands, execute an operation on them,
1393and produce a single value. The resulting value of the bitwise binary
1394operators is always the same type as its first operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1395</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1396<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1397<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1398Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1399<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1400<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1401<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 +0000[diff] [blame]1402</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1403<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1404<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1405its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1406<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1407<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1408 href="#t_integral">integral</a> values. Both arguments must have
1409identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1410<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1411<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1412<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1413<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1414<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1415 <tbody>
1416 <tr>
1417 <td>In0</td>
1418 <td>In1</td>
1419 <td>Out</td>
1420 </tr>
1421 <tr>
1422 <td>0</td>
1423 <td>0</td>
1424 <td>0</td>
1425 </tr>
1426 <tr>
1427 <td>0</td>
1428 <td>1</td>
1429 <td>0</td>
1430 </tr>
1431 <tr>
1432 <td>1</td>
1433 <td>0</td>
1434 <td>0</td>
1435 </tr>
1436 <tr>
1437 <td>1</td>
1438 <td>1</td>
1439 <td>1</td>
1440 </tr>
1441 </tbody>
1442</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1443</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1444<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1445<pre> &lt;result&gt; = and int 4, %var <i>; yields {int}:result = 4 &amp; %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1446 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1447 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1448</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1449</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1450<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1451<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1452<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1453<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1454<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 +0000[diff] [blame]1455</pre>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1456<h5>Overview:</h5>
1457<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1458or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1459<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1460<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1461 href="#t_integral">integral</a> values. Both arguments must have
1462identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1463<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1464<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1465<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1466<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1467<table border="1" cellspacing="0" cellpadding="4">
1468 <tbody>
1469 <tr>
1470 <td>In0</td>
1471 <td>In1</td>
1472 <td>Out</td>
1473 </tr>
1474 <tr>
1475 <td>0</td>
1476 <td>0</td>
1477 <td>0</td>
1478 </tr>
1479 <tr>
1480 <td>0</td>
1481 <td>1</td>
1482 <td>1</td>
1483 </tr>
1484 <tr>
1485 <td>1</td>
1486 <td>0</td>
1487 <td>1</td>
1488 </tr>
1489 <tr>
1490 <td>1</td>
1491 <td>1</td>
1492 <td>1</td>
1493 </tr>
1494 </tbody>
1495</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1496</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1497<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1498<pre> &lt;result&gt; = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1499 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1500 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1501</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1502</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1503<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1504<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1505Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1506<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1507<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1508<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 +0000[diff] [blame]1509</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1510<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1511<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1512or of its two operands. The <tt>xor</tt> is used to implement the
1513"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1514<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1515<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1516 href="#t_integral">integral</a> values. Both arguments must have
1517identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1518<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1519<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1520<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1521<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1522<table border="1" cellspacing="0" cellpadding="4">
1523 <tbody>
1524 <tr>
1525 <td>In0</td>
1526 <td>In1</td>
1527 <td>Out</td>
1528 </tr>
1529 <tr>
1530 <td>0</td>
1531 <td>0</td>
1532 <td>0</td>
1533 </tr>
1534 <tr>
1535 <td>0</td>
1536 <td>1</td>
1537 <td>1</td>
1538 </tr>
1539 <tr>
1540 <td>1</td>
1541 <td>0</td>
1542 <td>1</td>
1543 </tr>
1544 <tr>
1545 <td>1</td>
1546 <td>1</td>
1547 <td>0</td>
1548 </tr>
1549 </tbody>
1550</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1551</div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1552<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1553<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1554<pre> &lt;result&gt; = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1555 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
1556 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1557 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1558</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1559</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1560<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1561<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
1562Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1563<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1564<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1565<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 +0000[diff] [blame]1566</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1567<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1568<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
1569the left a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1570<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1571<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1572 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1573type.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1574<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1575<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1576<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1577<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 +0000[diff] [blame]1578 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1579 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1580</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1581</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1582<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1583<div class="doc_subsubsection"> <a name="i_shr">'<tt>shr</tt>'
1584Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1585<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1586<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1587<pre> &lt;result&gt; = shr &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1588</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1589<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1590<p>The '<tt>shr</tt>' instruction returns the first operand shifted to
1591the right a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1592<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1593<p>The first argument to the '<tt>shr</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1594 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1595type.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1596<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1597<p>If the first argument is a <a href="#t_signed">signed</a> type, the
1598most significant bit is duplicated in the newly free'd bit positions.
1599If the first argument is unsigned, zero bits shall fill the empty
1600positions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1601<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1602<pre> &lt;result&gt; = shr int 4, ubyte %var <i>; yields {int}:result = 4 &gt;&gt; %var</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +0000[diff] [blame]1603 &lt;result&gt; = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1604 &lt;result&gt; = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +0000[diff] [blame]1605 &lt;result&gt; = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1606 &lt;result&gt; = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1607</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1608</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1609<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1610<div class="doc_subsection"> <a name="memoryops">Memory Access
1611Operations</a></div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1612<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1613<p>A key design point of an SSA-based representation is how it
1614represents memory. In LLVM, no memory locations are in SSA form, which
1615makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1616allocate, and free memory in LLVM.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1617</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1618<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1619<div class="doc_subsubsection"> <a name="i_malloc">'<tt>malloc</tt>'
1620Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1621<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1622<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1623<pre> &lt;result&gt; = malloc &lt;type&gt;, uint &lt;NumElements&gt; <i>; yields {type*}:result</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1624 &lt;result&gt; = malloc &lt;type&gt; <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1625</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1626<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1627<p>The '<tt>malloc</tt>' instruction allocates memory from the system
1628heap and returns a pointer to it.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1629<h5>Arguments:</h5>
John Criswell6e4ca612004-02-24 16:13:56 +0000[diff] [blame]1630<p>The '<tt>malloc</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
1631bytes of memory from the operating system and returns a pointer of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1632appropriate type to the program. The second form of the instruction is
1633a shorter version of the first instruction that defaults to allocating
1634one element.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1635<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1636<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1637<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
1638a pointer is returned.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1639<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1640<pre> %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1641
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1642 %size = <a
1643 href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1644 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1645 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1646</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1647</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1648<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1649<div class="doc_subsubsection"> <a name="i_free">'<tt>free</tt>'
1650Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1651<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1652<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1653<pre> free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1654</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1655<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1656<p>The '<tt>free</tt>' instruction returns memory back to the unused
1657memory heap, to be reallocated in the future.</p>
1658<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1659<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1660<p>'<tt>value</tt>' shall be a pointer value that points to a value
1661that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
1662instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1663<h5>Semantics:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1664<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1665after this instruction executes.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1666<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1667<pre> %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1668 free [4 x ubyte]* %array
1669</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1670</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1671<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1672<div class="doc_subsubsection"> <a name="i_alloca">'<tt>alloca</tt>'
1673Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1674<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1675<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1676<pre> &lt;result&gt; = alloca &lt;type&gt;, uint &lt;NumElements&gt; <i>; yields {type*}:result</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1677 &lt;result&gt; = alloca &lt;type&gt; <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1678</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1679<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1680<p>The '<tt>alloca</tt>' instruction allocates memory on the current
1681stack frame of the procedure that is live until the current function
1682returns to its caller.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1683<h5>Arguments:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1684<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1685bytes of memory on the runtime stack, returning a pointer of the
1686appropriate type to the program. The second form of the instruction is
1687a shorter version of the first that defaults to allocating one element.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1688<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1689<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1690<p>Memory is allocated, a pointer is returned. '<tt>alloca</tt>'d
1691memory is automatically released when the function returns. The '<tt>alloca</tt>'
1692instruction is commonly used to represent automatic variables that must
1693have an address available. When the function returns (either with the <tt><a
1694 href="#i_ret">ret</a></tt> or <tt><a href="#i_invoke">invoke</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1695instructions), the memory is reclaimed.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1696<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1697<pre> %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1698 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1699</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1700</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1701<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1702<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
1703Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1704<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1705<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1706<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 +0000[diff] [blame]1707<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1708<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1709<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1710<p>The argument to the '<tt>load</tt>' instruction specifies the memory
1711address to load from. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +0000[diff] [blame]1712 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1713marked as <tt>volatile</tt> then the optimizer is not allowed to modify
1714the number or order of execution of this <tt>load</tt> with other
1715volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
1716instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1717<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1718<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1719<h5>Examples:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1720<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1721 <a
1722 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1723 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1724</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1725</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1726<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1727<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
1728Instruction</a> </div>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1729<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1730<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 +0000[diff] [blame]1731 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 +0000[diff] [blame]1732</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1733<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1734<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1735<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1736<p>There are two arguments to the '<tt>store</tt>' instruction: a value
1737to store and an address to store it into. The type of the '<tt>&lt;pointer&gt;</tt>'
1738operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
1739operand. If the <tt>store</tt> is marked as <tt>volatile</tt> then the
1740optimizer is not allowed to modify the number or order of execution of
1741this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
1742 href="#i_store">store</a></tt> instructions.</p>
1743<h5>Semantics:</h5>
1744<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
1745at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1746<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1747<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
1748 <a
1749 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1750 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
1751</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1752<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1753<div class="doc_subsubsection">
1754 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
1755</div>
1756
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1757<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1758<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1759<pre>
1760 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
1761</pre>
1762
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1763<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1764
1765<p>
1766The '<tt>getelementptr</tt>' instruction is used to get the address of a
1767subelement of an aggregate data structure.</p>
1768
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1769<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1770
1771<p>This instruction takes a list of integer constants that indicate what
1772elements of the aggregate object to index to. The actual types of the arguments
1773provided depend on the type of the first pointer argument. The
1774'<tt>getelementptr</tt>' instruction is used to index down through the type
1775levels of a structure. When indexing into a structure, only <tt>uint</tt>
1776integer constants are allowed. When indexing into an array or pointer
1777<tt>int</tt> and <tt>long</tt> indexes are allowed of any sign.</p>
1778
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1779<p>For example, let's consider a C code fragment and how it gets
1780compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1781
1782<pre>
1783 struct RT {
1784 char A;
1785 int B[10][20];
1786 char C;
1787 };
1788 struct ST {
1789 int X;
1790 double Y;
1791 struct RT Z;
1792 };
1793
1794 int *foo(struct ST *s) {
1795 return &amp;s[1].Z.B[5][13];
1796 }
1797</pre>
1798
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1799<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1800
1801<pre>
1802 %RT = type { sbyte, [10 x [20 x int]], sbyte }
1803 %ST = type { int, double, %RT }
1804
Brian Gaeke7283e7c2004-07-02 21:08:14 +0000[diff] [blame]1805 implementation
1806
1807 int* %foo(%ST* %s) {
1808 entry:
1809 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1810 ret int* %reg
1811 }
1812</pre>
1813
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1814<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1815
1816<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
Chris Lattnere53e5082004-06-03 22:57:15 +0000[diff] [blame]1817on the pointer type that is being index into. <a href="#t_pointer">Pointer</a>
1818and <a href="#t_array">array</a> types require <tt>uint</tt>, <tt>int</tt>,
1819<tt>ulong</tt>, or <tt>long</tt> values, and <a href="#t_struct">structure</a>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1820types require <tt>uint</tt> <b>constants</b>.</p>
1821
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1822<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1823type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
1824}</tt>' type, a structure. The second index indexes into the third element of
1825the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
1826sbyte }</tt>' type, another structure. The third index indexes into the second
1827element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
1828array. The two dimensions of the array are subscripted into, yielding an
1829'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction return a pointer
1830to this element, thus computing a value of '<tt>int*</tt>' type.</p>
1831
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1832<p>Note that it is perfectly legal to index partially through a
1833structure, returning a pointer to an inner element. Because of this,
1834the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1835
1836<pre>
1837 int* "foo"(%ST* %s) {
1838 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
1839 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
1840 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
1841 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
1842 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
1843 ret int* %t5
1844 }
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1845</pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1846<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]1847<pre>
1848 <i>; yields [12 x ubyte]*:aptr</i>
1849 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
1850</pre>
1851
1852</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1853<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1854<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1855<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]1856<p>The instructions in this category are the "miscellaneous"
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1857instructions, which defy better classification.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1858</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1859<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1860<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
1861Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1862<div class="doc_text">
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]1863<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1864<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]1865<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1866<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
1867the SSA graph representing the function.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]1868<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1869<p>The type of the incoming values are specified with the first type
1870field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
1871as arguments, with one pair for each predecessor basic block of the
1872current block. Only values of <a href="#t_firstclass">first class</a>
1873type may be used as the value arguments to the PHI node. Only labels
1874may be used as the label arguments.</p>
1875<p>There must be no non-phi instructions between the start of a basic
1876block and the PHI instructions: i.e. PHI instructions must be first in
1877a basic block.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]1878<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1879<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
1880value specified by the parameter, depending on which basic block we
1881came from in the last <a href="#terminators">terminator</a> instruction.</p>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1882<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1883<pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi uint [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add uint %indvar, 1<br> br label %Loop<br></pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1884</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1885
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1886<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1887<div class="doc_subsubsection">
1888 <a name="i_cast">'<tt>cast .. to</tt>' Instruction</a>
1889</div>
1890
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1891<div class="doc_text">
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1892
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1893<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1894
1895<pre>
1896 &lt;result&gt; = cast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1897</pre>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1898
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1899<h5>Overview:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1900
1901<p>
1902The '<tt>cast</tt>' instruction is used as the primitive means to convert
1903integers to floating point, change data type sizes, and break type safety (by
1904casting pointers).
1905</p>
1906
1907
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1908<h5>Arguments:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1909
1910<p>
1911The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
1912class value, and a type to cast it to, which must also be a <a
1913href="#t_firstclass">first class</a> type.
1914</p>
1915
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1916<h5>Semantics:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1917
1918<p>
1919This instruction follows the C rules for explicit casts when determining how the
1920data being cast must change to fit in its new container.
1921</p>
1922
1923<p>
1924When casting to bool, any value that would be considered true in the context of
1925a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
1926all else are '<tt>false</tt>'.
1927</p>
1928
1929<p>
1930When extending an integral value from a type of one signness to another (for
1931example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
1932<b>source</b> value is signed, and zero-extended if the source value is
1933unsigned. <tt>bool</tt> values are always zero extended into either zero or
1934one.
1935</p>
1936
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]1937<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1938
1939<pre>
1940 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
Chris Lattner7bae3952002-06-25 18:03:17 +0000[diff] [blame]1941 %Y = cast int 123 to bool <i>; yields bool:true</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]1942</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1943</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]1944
1945<!-- _______________________________________________________________________ -->
1946<div class="doc_subsubsection">
1947 <a name="i_select">'<tt>select</tt>' Instruction</a>
1948</div>
1949
1950<div class="doc_text">
1951
1952<h5>Syntax:</h5>
1953
1954<pre>
1955 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
1956</pre>
1957
1958<h5>Overview:</h5>
1959
1960<p>
1961The '<tt>select</tt>' instruction is used to choose one value based on a
1962condition, without branching.
1963</p>
1964
1965
1966<h5>Arguments:</h5>
1967
1968<p>
1969The '<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.
1970</p>
1971
1972<h5>Semantics:</h5>
1973
1974<p>
1975If the boolean condition evaluates to true, the instruction returns the first
1976value argument, otherwise it returns the second value argument.
1977</p>
1978
1979<h5>Example:</h5>
1980
1981<pre>
1982 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
1983</pre>
1984</div>
1985
1986
1987
1988
1989
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]1990<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1991<div class="doc_subsubsection"> <a name="i_call">'<tt>call</tt>'
1992Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1993<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1994<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1995<pre> &lt;result&gt; = call &lt;ty&gt;* &lt;fnptrval&gt;(&lt;param list&gt;)<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1996<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1997<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1998<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1999<p>This instruction requires several arguments:</p>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2000<ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2001 <li>
2002 <p>'<tt>ty</tt>': shall be the signature of the pointer to function
2003value being invoked. The argument types must match the types implied
2004by this signature.</p>
2005 </li>
2006 <li>
2007 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a
2008function to be invoked. In most cases, this is a direct function
2009invocation, but indirect <tt>call</tt>s are just as possible,
2010calling an arbitrary pointer to function values.</p>
2011 </li>
2012 <li>
2013 <p>'<tt>function args</tt>': argument list whose types match the
2014function signature argument types. If the function signature
2015indicates the function accepts a variable number of arguments, the
2016extra arguments can be specified.</p>
2017 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2018</ol>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2019<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2020<p>The '<tt>call</tt>' instruction is used to cause control flow to
2021transfer to a specified function, with its incoming arguments bound to
2022the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
2023instruction in the called function, control flow continues with the
2024instruction after the function call, and the return value of the
2025function is bound to the result argument. This is a simpler case of
2026the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2027<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2028<pre> %retval = call int %test(int %argc)<br> call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);<br></pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2029</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2030
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2031<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2032<div class="doc_subsubsection">
2033 <a name="i_vanext">'<tt>vanext</tt>' Instruction</a>
2034</div>
2035
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2036<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2037
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2038<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2039
2040<pre>
2041 &lt;resultarglist&gt; = vanext &lt;va_list&gt; &lt;arglist&gt;, &lt;argty&gt;
2042</pre>
2043
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2044<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2045
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2046<p>The '<tt>vanext</tt>' instruction is used to access arguments passed
2047through the "variable argument" area of a function call. It is used to
2048implement the <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2049
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2050<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2051
2052<p>This instruction takes a <tt>va_list</tt> value and the type of the
2053argument. It returns another <tt>va_list</tt>. The actual type of
2054<tt>va_list</tt> may be defined differently for different targets. Most targets
2055use a <tt>va_list</tt> type of <tt>sbyte*</tt> or some other pointer type.</p>
2056
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2057<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2058
2059<p>The '<tt>vanext</tt>' instruction advances the specified <tt>va_list</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2060past an argument of the specified type. In conjunction with the <a
2061 href="#i_vaarg"><tt>vaarg</tt></a> instruction, it is used to implement
2062the <tt>va_arg</tt> macro available in C. For more information, see
2063the variable argument handling <a href="#int_varargs">Intrinsic
2064Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2065
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2066<p>It is legal for this instruction to be called in a function which
2067does not take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2068function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2069
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2070<p><tt>vanext</tt> is an LLVM instruction instead of an <a
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2071href="#intrinsics">intrinsic function</a> because it takes a type as an
2072argument. The type refers to the current argument in the <tt>va_list</tt>, it
2073tells the compiler how far on the stack it needs to advance to find the next
2074argument</p>
2075
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2076<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2077
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2078<p>See the <a href="#int_varargs">variable argument processing</a>
2079section.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2080
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2081</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2082
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2083<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2084<div class="doc_subsubsection">
2085 <a name="i_vaarg">'<tt>vaarg</tt>' Instruction</a>
2086</div>
2087
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2088<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2089
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2090<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2091
2092<pre>
2093 &lt;resultval&gt; = vaarg &lt;va_list&gt; &lt;arglist&gt;, &lt;argty&gt;
2094</pre>
2095
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2096<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2097
2098<p>The '<tt>vaarg</tt>' instruction is used to access arguments passed through
2099the "variable argument" area of a function call. It is used to implement the
2100<tt>va_arg</tt> macro in C.</p>
2101
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2102<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2103
2104<p>This instruction takes a <tt>va_list</tt> value and the type of the
2105argument. It returns a value of the specified argument type. Again, the actual
2106type of <tt>va_list</tt> is target specific.</p>
2107
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2108<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2109
2110<p>The '<tt>vaarg</tt>' instruction loads an argument of the specified type from
2111the specified <tt>va_list</tt>. In conjunction with the <a
2112href="#i_vanext"><tt>vanext</tt></a> instruction, it is used to implement the
2113<tt>va_arg</tt> macro available in C. For more information, see the variable
2114argument handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
2115
2116<p>It is legal for this instruction to be called in a function which does not
2117take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2118function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2119
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2120<p><tt>vaarg</tt> is an LLVM instruction instead of an <a
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2121href="#intrinsics">intrinsic function</a> because it takes an type as an
2122argument.</p>
2123
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2124<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2125
2126<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
2127
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2128</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2129
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2130<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2131<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
2132<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2133
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2134<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2135
2136<p>LLVM supports the notion of an "intrinsic function". These functions have
2137well known names and semantics, and are required to follow certain
2138restrictions. Overall, these instructions represent an extension mechanism for
2139the LLVM language that does not require changing all of the transformations in
2140LLVM to add to the language (or the bytecode reader/writer, the parser,
2141etc...).</p>
2142
2143<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix, this
2144prefix is reserved in LLVM for intrinsic names, thus functions may not be named
2145this. Intrinsic functions must always be external functions: you cannot define
2146the body of intrinsic functions. Intrinsic functions may only be used in call
2147or invoke instructions: it is illegal to take the address of an intrinsic
2148function. Additionally, because intrinsic functions are part of the LLVM
2149language, it is required that they all be documented here if any are added.</p>
2150
2151
2152<p>
2153Adding an intrinsic to LLVM is straight-forward if it is possible to express the
2154concept in LLVM directly (ie, code generator support is not _required_). To do
2155this, extend the default implementation of the IntrinsicLowering class to handle
2156the intrinsic. Code generators use this class to lower intrinsics they do not
2157understand to raw LLVM instructions that they do.
2158</p>
2159
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2160</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2161
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2162<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2163<div class="doc_subsection">
2164 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
2165</div>
2166
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2167<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2168
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2169<p>Variable argument support is defined in LLVM with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2170 href="#i_vanext"><tt>vanext</tt></a> instruction and these three
2171intrinsic functions. These functions are related to the similarly
2172named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2173
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2174<p>All of these functions operate on arguments that use a
2175target-specific value type "<tt>va_list</tt>". The LLVM assembly
2176language reference manual does not define what this type is, so all
2177transformations should be prepared to handle intrinsics with any type
2178used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2179
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2180<p>This example shows how the <a href="#i_vanext"><tt>vanext</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2181instruction and the variable argument handling intrinsic functions are
2182used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2183
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2184<pre>
2185int %test(int %X, ...) {
2186 ; Initialize variable argument processing
2187 %ap = call sbyte* %<a href="#i_va_start">llvm.va_start</a>()
2188
2189 ; Read a single integer argument
2190 %tmp = vaarg sbyte* %ap, int
2191
2192 ; Advance to the next argument
2193 %ap2 = vanext sbyte* %ap, int
2194
2195 ; Demonstrate usage of llvm.va_copy and llvm.va_end
2196 %aq = call sbyte* %<a href="#i_va_copy">llvm.va_copy</a>(sbyte* %ap2)
2197 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte* %aq)
2198
2199 ; Stop processing of arguments.
2200 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte* %ap2)
2201 ret int %tmp
2202}
2203</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2204</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2205
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2206<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2207<div class="doc_subsubsection">
2208 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
2209</div>
2210
2211
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2212<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2213<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2214<pre> call &lt;va_list&gt; ()* %llvm.va_start()<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2215<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2216<p>The '<tt>llvm.va_start</tt>' intrinsic returns a new <tt>&lt;arglist&gt;</tt>
2217for subsequent use by the variable argument intrinsics.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2218<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2219<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2220macro available in C. In a target-dependent way, it initializes and
2221returns a <tt>va_list</tt> element, so that the next <tt>vaarg</tt>
2222will produce the first variable argument passed to the function. Unlike
2223the C <tt>va_start</tt> macro, this intrinsic does not need to know the
2224last argument of the function, the compiler can figure that out.</p>
2225<p>Note that this intrinsic function is only legal to be called from
2226within the body of a variable argument function.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2227</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2228
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2229<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2230<div class="doc_subsubsection">
2231 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
2232</div>
2233
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2234<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2235<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2236<pre> call void (&lt;va_list&gt;)* %llvm.va_end(&lt;va_list&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2237<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2238<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
2239which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
2240or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2241<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2242<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2243<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2244<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2245macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
2246Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
2247 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
2248with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2249</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2250
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2251<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2252<div class="doc_subsubsection">
2253 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
2254</div>
2255
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2256<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2257
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2258<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2259
2260<pre>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2261 call &lt;va_list&gt; (&lt;va_list&gt;)* %llvm.va_copy(&lt;va_list&gt; &lt;destarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2262</pre>
2263
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2264<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2265
2266<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
2267from the source argument list to the destination argument list.</p>
2268
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2269<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2270
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2271<p>The argument is the <tt>va_list</tt> to copy.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2272
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2273<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2274
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2275<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2276macro available in C. In a target-dependent way, it copies the source
2277<tt>va_list</tt> element into the returned list. This intrinsic is necessary
Chris Lattnerfcd37252004-06-21 22:52:48 +0000[diff] [blame]2278because the <tt><a href="#i_va_start">llvm.va_start</a></tt> intrinsic may be
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2279arbitrarily complex and require memory allocation, for example.</p>
2280
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2281</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2282
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2283<!-- ======================================================================= -->
2284<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2285 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
2286</div>
2287
2288<div class="doc_text">
2289
2290<p>
2291LLVM support for <a href="GarbageCollection.html">Accurate Garbage
2292Collection</a> requires the implementation and generation of these intrinsics.
2293These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
2294stack</a>, as well as garbage collector implementations that require <a
2295href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
2296Front-ends for type-safe garbage collected languages should generate these
2297intrinsics to make use of the LLVM garbage collectors. For more details, see <a
2298href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
2299</p>
2300</div>
2301
2302<!-- _______________________________________________________________________ -->
2303<div class="doc_subsubsection">
2304 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
2305</div>
2306
2307<div class="doc_text">
2308
2309<h5>Syntax:</h5>
2310
2311<pre>
2312 call void (&lt;ty&gt;**, &lt;ty2&gt;*)* %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
2313</pre>
2314
2315<h5>Overview:</h5>
2316
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2317<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2318the code generator, and allows some metadata to be associated with it.</p>
2319
2320<h5>Arguments:</h5>
2321
2322<p>The first argument specifies the address of a stack object that contains the
2323root pointer. The second pointer (which must be either a constant or a global
2324value address) contains the meta-data to be associated with the root.</p>
2325
2326<h5>Semantics:</h5>
2327
2328<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
2329location. At compile-time, the code generator generates information to allow
2330the runtime to find the pointer at GC safe points.
2331</p>
2332
2333</div>
2334
2335
2336<!-- _______________________________________________________________________ -->
2337<div class="doc_subsubsection">
2338 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
2339</div>
2340
2341<div class="doc_text">
2342
2343<h5>Syntax:</h5>
2344
2345<pre>
2346 call sbyte* (sbyte**)* %llvm.gcread(sbyte** %Ptr)
2347</pre>
2348
2349<h5>Overview:</h5>
2350
2351<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
2352locations, allowing garbage collector implementations that require read
2353barriers.</p>
2354
2355<h5>Arguments:</h5>
2356
2357<p>The argument is the address to read from, which should be an address
2358allocated from the garbage collector.</p>
2359
2360<h5>Semantics:</h5>
2361
2362<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
2363instruction, but may be replaced with substantially more complex code by the
2364garbage collector runtime, as needed.</p>
2365
2366</div>
2367
2368
2369<!-- _______________________________________________________________________ -->
2370<div class="doc_subsubsection">
2371 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
2372</div>
2373
2374<div class="doc_text">
2375
2376<h5>Syntax:</h5>
2377
2378<pre>
2379 call void (sbyte*, sbyte**)* %llvm.gcwrite(sbyte* %P1, sbyte** %P2)
2380</pre>
2381
2382<h5>Overview:</h5>
2383
2384<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
2385locations, allowing garbage collector implementations that require write
2386barriers (such as generational or reference counting collectors).</p>
2387
2388<h5>Arguments:</h5>
2389
2390<p>The first argument is the reference to store, and the second is the heap
2391location to store to.</p>
2392
2393<h5>Semantics:</h5>
2394
2395<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
2396instruction, but may be replaced with substantially more complex code by the
2397garbage collector runtime, as needed.</p>
2398
2399</div>
2400
2401
2402
2403<!-- ======================================================================= -->
2404<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2405 <a name="int_codegen">Code Generator Intrinsics</a>
2406</div>
2407
2408<div class="doc_text">
2409<p>
2410These intrinsics are provided by LLVM to expose special features that may only
2411be implemented with code generator support.
2412</p>
2413
2414</div>
2415
2416<!-- _______________________________________________________________________ -->
2417<div class="doc_subsubsection">
2418 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
2419</div>
2420
2421<div class="doc_text">
2422
2423<h5>Syntax:</h5>
2424<pre>
2425 call void* ()* %llvm.returnaddress(uint &lt;level&gt;)
2426</pre>
2427
2428<h5>Overview:</h5>
2429
2430<p>
2431The '<tt>llvm.returnaddress</tt>' intrinsic returns a target-specific value
2432indicating the return address of the current function or one of its callers.
2433</p>
2434
2435<h5>Arguments:</h5>
2436
2437<p>
2438The argument to this intrinsic indicates which function to return the address
2439for. Zero indicates the calling function, one indicates its caller, etc. The
2440argument is <b>required</b> to be a constant integer value.
2441</p>
2442
2443<h5>Semantics:</h5>
2444
2445<p>
2446The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
2447the return address of the specified call frame, or zero if it cannot be
2448identified. The value returned by this intrinsic is likely to be incorrect or 0
2449for arguments other than zero, so it should only be used for debugging purposes.
2450</p>
2451
2452<p>
2453Note that calling this intrinsic does not prevent function inlining or other
2454aggressive transformations, so the value returned may not that of the obvious
2455source-language caller.
2456</p>
2457</div>
2458
2459
2460<!-- _______________________________________________________________________ -->
2461<div class="doc_subsubsection">
2462 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
2463</div>
2464
2465<div class="doc_text">
2466
2467<h5>Syntax:</h5>
2468<pre>
2469 call void* ()* %llvm.frameaddress(uint &lt;level&gt;)
2470</pre>
2471
2472<h5>Overview:</h5>
2473
2474<p>
2475The '<tt>llvm.frameaddress</tt>' intrinsic returns the target-specific frame
2476pointer value for the specified stack frame.
2477</p>
2478
2479<h5>Arguments:</h5>
2480
2481<p>
2482The argument to this intrinsic indicates which function to return the frame
2483pointer for. Zero indicates the calling function, one indicates its caller,
2484etc. The argument is <b>required</b> to be a constant integer value.
2485</p>
2486
2487<h5>Semantics:</h5>
2488
2489<p>
2490The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
2491the frame address of the specified call frame, or zero if it cannot be
2492identified. The value returned by this intrinsic is likely to be incorrect or 0
2493for arguments other than zero, so it should only be used for debugging purposes.
2494</p>
2495
2496<p>
2497Note that calling this intrinsic does not prevent function inlining or other
2498aggressive transformations, so the value returned may not that of the obvious
2499source-language caller.
2500</p>
2501</div>
2502
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]2503<!-- _______________________________________________________________________ -->
2504<div class="doc_subsubsection">
2505 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
2506</div>
2507
2508<div class="doc_text">
2509
2510<h5>Syntax:</h5>
2511<pre>
2512 call void (sbyte *, uint, uint)* %llvm.prefetch(sbyte * &lt;address&gt;,
2513 uint &lt;rw&gt;,
2514 uint &lt;locality&gt;)
2515</pre>
2516
2517<h5>Overview:</h5>
2518
2519
2520<p>
2521The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
2522a prefetch instruction if supported, otherwise it is a noop. Prefetches have no
Chris Lattner2a615362005-02-28 19:47:14 +0000[diff] [blame^]2523effect on the behavior of the program, but can change its performance
2524characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]2525</p>
2526
2527<h5>Arguments:</h5>
2528
2529<p>
2530<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
2531determining if the fetch should be for a read (0) or write (1), and
2532<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
2533locality, to (3) - exteremely local keep in cache. The <tt>rw</tt> and
2534<tt>locality</tt> arguments must be constant integers.
2535</p>
2536
2537<h5>Semantics:</h5>
2538
2539<p>
2540This intrinsic does not modify the behavior of the program. In particular,
2541prefetches cannot trap and do not produce a value. On targets that support this
2542intrinsic, the prefetch can provide hints to the processor cache for better
2543performance.
2544</p>
2545
2546</div>
2547
2548
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]2549<!-- ======================================================================= -->
2550<div class="doc_subsection">
2551 <a name="int_os">Operating System Intrinsics</a>
2552</div>
2553
2554<div class="doc_text">
2555<p>
2556These intrinsics are provided by LLVM to support the implementation of
2557operating system level code.
2558</p>
2559
2560</div>
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2561
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2562<!-- _______________________________________________________________________ -->
2563<div class="doc_subsubsection">
2564 <a name="i_readport">'<tt>llvm.readport</tt>' Intrinsic</a>
2565</div>
2566
2567<div class="doc_text">
2568
2569<h5>Syntax:</h5>
2570<pre>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]2571 call &lt;integer type&gt; (&lt;integer type&gt;)* %llvm.readport (&lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2572</pre>
2573
2574<h5>Overview:</h5>
2575
2576<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]2577The '<tt>llvm.readport</tt>' intrinsic reads data from the specified hardware
2578I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2579</p>
2580
2581<h5>Arguments:</h5>
2582
2583<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]2584The argument to this intrinsic indicates the hardware I/O address from which
2585to read the data. The address is in the hardware I/O address namespace (as
2586opposed to being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2587</p>
2588
2589<h5>Semantics:</h5>
2590
2591<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]2592The '<tt>llvm.readport</tt>' intrinsic reads data from the hardware I/O port
2593specified by <i>address</i> and returns the value. The address and return
2594value must be integers, but the size is dependent upon the platform upon which
2595the program is code generated. For example, on x86, the address must be an
2596unsigned 16 bit value, and the return value must be 8, 16, or 32 bits.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2597</p>
2598
2599</div>
2600
2601<!-- _______________________________________________________________________ -->
2602<div class="doc_subsubsection">
2603 <a name="i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a>
2604</div>
2605
2606<div class="doc_text">
2607
2608<h5>Syntax:</h5>
2609<pre>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2610 call void (&lt;integer type&gt;, &lt;integer type&gt;)*
2611 %llvm.writeport (&lt;integer type&gt; &lt;value&gt;,
2612 &lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2613</pre>
2614
2615<h5>Overview:</h5>
2616
2617<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]2618The '<tt>llvm.writeport</tt>' intrinsic writes data to the specified hardware
2619I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2620</p>
2621
2622<h5>Arguments:</h5>
2623
2624<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2625The first argument is the value to write to the I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2626</p>
2627
2628<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2629The second argument indicates the hardware I/O address to which data should be
2630written. The address is in the hardware I/O address namespace (as opposed to
2631being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2632</p>
2633
2634<h5>Semantics:</h5>
2635
2636<p>
2637The '<tt>llvm.writeport</tt>' intrinsic writes <i>value</i> to the I/O port
2638specified by <i>address</i>. The address and value must be integers, but the
2639size is dependent upon the platform upon which the program is code generated.
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]2640For example, on x86, the address must be an unsigned 16 bit value, and the
2641value written must be 8, 16, or 32 bits in length.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]2642</p>
2643
2644</div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2645
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2646<!-- _______________________________________________________________________ -->
2647<div class="doc_subsubsection">
2648 <a name="i_readio">'<tt>llvm.readio</tt>' Intrinsic</a>
2649</div>
2650
2651<div class="doc_text">
2652
2653<h5>Syntax:</h5>
2654<pre>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2655 call &lt;result&gt; (&lt;ty&gt;*)* %llvm.readio (&lt;ty&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2656</pre>
2657
2658<h5>Overview:</h5>
2659
2660<p>
2661The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
2662address.
2663</p>
2664
2665<h5>Arguments:</h5>
2666
2667<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2668The argument to this intrinsic is a pointer indicating the memory address from
2669which to read the data. The data must be a
2670<a href="#t_firstclass">first class</a> type.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2671</p>
2672
2673<h5>Semantics:</h5>
2674
2675<p>
2676The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2677location specified by <i>pointer</i> and returns the value. The argument must
2678be a pointer, and the return value must be a
2679<a href="#t_firstclass">first class</a> type. However, certain architectures
2680may not support I/O on all first class types. For example, 32 bit processors
2681may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2682</p>
2683
2684<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2685This intrinsic enforces an in-order memory model for llvm.readio and
2686llvm.writeio calls on machines that use dynamic scheduling. Dynamically
2687scheduled processors may execute loads and stores out of order, re-ordering at
2688run time accesses to memory mapped I/O registers. Using these intrinsics
2689ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2690</p>
2691
2692</div>
2693
2694<!-- _______________________________________________________________________ -->
2695<div class="doc_subsubsection">
2696 <a name="i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a>
2697</div>
2698
2699<div class="doc_text">
2700
2701<h5>Syntax:</h5>
2702<pre>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2703 call void (&lt;ty1&gt;, &lt;ty2&gt;*)* %llvm.writeio (&lt;ty1&gt; &lt;value&gt;, &lt;ty2&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2704</pre>
2705
2706<h5>Overview:</h5>
2707
2708<p>
2709The '<tt>llvm.writeio</tt>' intrinsic writes data to the specified memory
2710mapped I/O address.
2711</p>
2712
2713<h5>Arguments:</h5>
2714
2715<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2716The first argument is the value to write to the memory mapped I/O location.
2717The second argument is a pointer indicating the memory address to which the
2718data should be written.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2719</p>
2720
2721<h5>Semantics:</h5>
2722
2723<p>
2724The '<tt>llvm.writeio</tt>' intrinsic writes <i>value</i> to the memory mapped
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2725I/O address specified by <i>pointer</i>. The value must be a
2726<a href="#t_firstclass">first class</a> type. However, certain architectures
2727may not support I/O on all first class types. For example, 32 bit processors
2728may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2729</p>
2730
2731<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]2732This intrinsic enforces an in-order memory model for llvm.readio and
2733llvm.writeio calls on machines that use dynamic scheduling. Dynamically
2734scheduled processors may execute loads and stores out of order, re-ordering at
2735run time accesses to memory mapped I/O registers. Using these intrinsics
2736ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]2737</p>
2738
2739</div>
2740
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2741<!-- ======================================================================= -->
2742<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2743 <a name="int_libc">Standard C Library Intrinsics</a>
2744</div>
2745
2746<div class="doc_text">
2747<p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2748LLVM provides intrinsics for a few important standard C library functions.
2749These intrinsics allow source-language front-ends to pass information about the
2750alignment of the pointer arguments to the code generator, providing opportunity
2751for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2752</p>
2753
2754</div>
2755
2756<!-- _______________________________________________________________________ -->
2757<div class="doc_subsubsection">
2758 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
2759</div>
2760
2761<div class="doc_text">
2762
2763<h5>Syntax:</h5>
2764<pre>
2765 call void (sbyte*, sbyte*, uint, uint)* %llvm.memcpy(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
2766 uint &lt;len&gt;, uint &lt;align&gt;)
2767</pre>
2768
2769<h5>Overview:</h5>
2770
2771<p>
2772The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
2773location to the destination location.
2774</p>
2775
2776<p>
2777Note that, unlike the standard libc function, the <tt>llvm.memcpy</tt> intrinsic
2778does not return a value, and takes an extra alignment argument.
2779</p>
2780
2781<h5>Arguments:</h5>
2782
2783<p>
2784The first argument is a pointer to the destination, the second is a pointer to
2785the source. The third argument is an (arbitrarily sized) integer argument
2786specifying the number of bytes to copy, and the fourth argument is the alignment
2787of the source and destination locations.
2788</p>
2789
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]2790<p>
2791If the call to this intrinisic has an alignment value that is not 0 or 1, then
2792the caller guarantees that the size of the copy is a multiple of the alignment
2793and that both the source and destination pointers are aligned to that boundary.
2794</p>
2795
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2796<h5>Semantics:</h5>
2797
2798<p>
2799The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
2800location to the destination location, which are not allowed to overlap. It
2801copies "len" bytes of memory over. If the argument is known to be aligned to
2802some boundary, this can be specified as the fourth argument, otherwise it should
2803be set to 0 or 1.
2804</p>
2805</div>
2806
2807
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]2808<!-- _______________________________________________________________________ -->
2809<div class="doc_subsubsection">
2810 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
2811</div>
2812
2813<div class="doc_text">
2814
2815<h5>Syntax:</h5>
2816<pre>
2817 call void (sbyte*, sbyte*, uint, uint)* %llvm.memmove(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
2818 uint &lt;len&gt;, uint &lt;align&gt;)
2819</pre>
2820
2821<h5>Overview:</h5>
2822
2823<p>
2824The '<tt>llvm.memmove</tt>' intrinsic moves a block of memory from the source
2825location to the destination location. It is similar to the '<tt>llvm.memcpy</tt>'
2826intrinsic but allows the two memory locations to overlap.
2827</p>
2828
2829<p>
2830Note that, unlike the standard libc function, the <tt>llvm.memmove</tt> intrinsic
2831does not return a value, and takes an extra alignment argument.
2832</p>
2833
2834<h5>Arguments:</h5>
2835
2836<p>
2837The first argument is a pointer to the destination, the second is a pointer to
2838the source. The third argument is an (arbitrarily sized) integer argument
2839specifying the number of bytes to copy, and the fourth argument is the alignment
2840of the source and destination locations.
2841</p>
2842
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]2843<p>
2844If the call to this intrinisic has an alignment value that is not 0 or 1, then
2845the caller guarantees that the size of the copy is a multiple of the alignment
2846and that both the source and destination pointers are aligned to that boundary.
2847</p>
2848
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]2849<h5>Semantics:</h5>
2850
2851<p>
2852The '<tt>llvm.memmove</tt>' intrinsic copies a block of memory from the source
2853location to the destination location, which may overlap. It
2854copies "len" bytes of memory over. If the argument is known to be aligned to
2855some boundary, this can be specified as the fourth argument, otherwise it should
2856be set to 0 or 1.
2857</p>
2858</div>
2859
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2860
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2861<!-- _______________________________________________________________________ -->
2862<div class="doc_subsubsection">
2863 <a name="i_memset">'<tt>llvm.memset</tt>' Intrinsic</a>
2864</div>
2865
2866<div class="doc_text">
2867
2868<h5>Syntax:</h5>
2869<pre>
2870 call void (sbyte*, ubyte, uint, uint)* %llvm.memset(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
2871 uint &lt;len&gt;, uint &lt;align&gt;)
2872</pre>
2873
2874<h5>Overview:</h5>
2875
2876<p>
2877The '<tt>llvm.memset</tt>' intrinsic fills a block of memory with a particular
2878byte value.
2879</p>
2880
2881<p>
2882Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
2883does not return a value, and takes an extra alignment argument.
2884</p>
2885
2886<h5>Arguments:</h5>
2887
2888<p>
2889The first argument is a pointer to the destination to fill, the second is the
2890byte value to fill it with, the third argument is an (arbitrarily sized) integer
2891argument specifying the number of bytes to fill, and the fourth argument is the
2892known alignment of destination location.
2893</p>
2894
2895<p>
2896If the call to this intrinisic has an alignment value that is not 0 or 1, then
2897the caller guarantees that the size of the copy is a multiple of the alignment
2898and that the destination pointer is aligned to that boundary.
2899</p>
2900
2901<h5>Semantics:</h5>
2902
2903<p>
2904The '<tt>llvm.memset</tt>' intrinsic fills "len" bytes of memory starting at the
2905destination location. If the argument is known to be aligned to some boundary,
2906this can be specified as the fourth argument, otherwise it should be set to 0 or
29071.
2908</p>
2909</div>
2910
2911
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]2912<!-- _______________________________________________________________________ -->
2913<div class="doc_subsubsection">
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +0000[diff] [blame]2914 <a name="i_isunordered">'<tt>llvm.isunordered</tt>' Intrinsic</a>
2915</div>
2916
2917<div class="doc_text">
2918
2919<h5>Syntax:</h5>
2920<pre>
2921 call bool (&lt;float or double&gt;, &lt;float or double&gt;)* %llvm.isunordered(&lt;float or double&gt; Val1,
2922 &lt;float or double&gt; Val2)
2923</pre>
2924
2925<h5>Overview:</h5>
2926
2927<p>
2928The '<tt>llvm.isunordered</tt>' intrinsic returns true if either or both of the
2929specified floating point values is a NAN.
2930</p>
2931
2932<h5>Arguments:</h5>
2933
2934<p>
2935The arguments are floating point numbers of the same type.
2936</p>
2937
2938<h5>Semantics:</h5>
2939
2940<p>
2941If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
2942false.
2943</p>
2944</div>
2945
2946
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]2947
2948
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2949<!-- ======================================================================= -->
2950<div class="doc_subsection">
2951 <a name="int_debugger">Debugger Intrinsics</a>
2952</div>
2953
2954<div class="doc_text">
2955<p>
2956The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
2957are described in the <a
2958href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
2959Debugging</a> document.
2960</p>
2961</div>
2962
2963
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2964<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2965<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2966<address>
2967 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
2968 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
2969 <a href="http://validator.w3.org/check/referer"><img
2970 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
2971
2972 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
2973 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
2974 Last modified: $Date$
2975</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2976</body>
2977</html>