Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 4e9aba7c121bddc51ecbe2fa29582731a474c30f / . / docs / LangRef.html
blob: b7c59a54146b9294a0ac95b9cc912c7d9d8c6b96 [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 Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]24 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]25 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame^]26 <li><a href="#functionstructure">Functions</a></li>
27 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]28 </ol>
29 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]30 <li><a href="#typesystem">Type System</a>
31 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]32 <li><a href="#t_primitive">Primitive Types</a>
33 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]34 <li><a href="#t_classifications">Type Classifications</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]35 </ol>
36 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]37 <li><a href="#t_derived">Derived Types</a>
38 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]39 <li><a href="#t_array">Array Type</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]40 <li><a href="#t_function">Function Type</a></li>
41 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]42 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]43 <li><a href="#t_packed">Packed Type</a></li>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]44 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]45 </ol>
46 </li>
47 </ol>
48 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]49 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]50 <ol>
51 <li><a href="#simpleconstants">Simple Constants</a>
52 <li><a href="#aggregateconstants">Aggregate Constants</a>
53 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
54 <li><a href="#undefvalues">Undefined Values</a>
55 <li><a href="#constantexprs">Constant Expressions</a>
56 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]57 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]58 <li><a href="#instref">Instruction Reference</a>
59 <ol>
60 <li><a href="#terminators">Terminator Instructions</a>
61 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]62 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
63 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]64 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
65 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]66 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]67 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]68 </ol>
69 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]70 <li><a href="#binaryops">Binary Operations</a>
71 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]72 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
73 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
74 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
75 <li><a href="#i_div">'<tt>div</tt>' Instruction</a></li>
76 <li><a href="#i_rem">'<tt>rem</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]77 <li><a href="#i_setcc">'<tt>set<i>cc</i></tt>' Instructions</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]78 </ol>
79 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]80 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
81 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]82 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]83 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]84 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
85 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
86 <li><a href="#i_shr">'<tt>shr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]87 </ol>
88 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]89 <li><a href="#memoryops">Memory Access Operations</a>
90 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]91 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
92 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
93 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
94 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
95 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
96 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
97 </ol>
98 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]99 <li><a href="#otherops">Other Operations</a>
100 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]101 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]102 <li><a href="#i_cast">'<tt>cast .. to</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]103 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Robert Bocchino3a558662006-01-05 17:37:02 +0000[diff] [blame]104 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]105 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]106 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]107 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]108 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]109 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]110 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]111 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]112 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]113 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]114 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
115 <ol>
116 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
117 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
118 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
119 </ol>
120 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]121 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
122 <ol>
123 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
124 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
125 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
126 </ol>
127 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]128 <li><a href="#int_codegen">Code Generator Intrinsics</a>
129 <ol>
130 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
131 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]132 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
133 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]134 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]135 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]136 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]137 </ol>
138 </li>
139 <li><a href="#int_os">Operating System Intrinsics</a>
140 <ol>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]141 <li><a href="#i_readport">'<tt>llvm.readport</tt>' Intrinsic</a></li>
142 <li><a href="#i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a></li>
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]143 <li><a href="#i_readio">'<tt>llvm.readio</tt>' Intrinsic</a></li>
144 <li><a href="#i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]145 </ol>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]146 <li><a href="#int_libc">Standard C Library Intrinsics</a>
147 <ol>
148 <li><a href="#i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a></li>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]149 <li><a href="#i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a></li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]150 <li><a href="#i_memset">'<tt>llvm.memset</tt>' Intrinsic</a></li>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]151 <li><a href="#i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a></li>
152 <li><a href="#i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]153
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]154 </ol>
155 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]156 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]157 <ol>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]158 <li><a href="#i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]159 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
160 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
161 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]162 </ol>
163 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]164 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]165 </ol>
166 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]167</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]168
169<div class="doc_author">
170 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
171 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]172</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]173
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]174<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]175<div class="doc_section"> <a name="abstract">Abstract </a></div>
176<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]177
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]178<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]179<p>This document is a reference manual for the LLVM assembly language.
180LLVM is an SSA based representation that provides type safety,
181low-level operations, flexibility, and the capability of representing
182'all' high-level languages cleanly. It is the common code
183representation used throughout all phases of the LLVM compilation
184strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]185</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]186
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]187<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]188<div class="doc_section"> <a name="introduction">Introduction</a> </div>
189<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]190
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]191<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]192
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]193<p>The LLVM code representation is designed to be used in three
194different forms: as an in-memory compiler IR, as an on-disk bytecode
195representation (suitable for fast loading by a Just-In-Time compiler),
196and as a human readable assembly language representation. This allows
197LLVM to provide a powerful intermediate representation for efficient
198compiler transformations and analysis, while providing a natural means
199to debug and visualize the transformations. The three different forms
200of LLVM are all equivalent. This document describes the human readable
201representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]202
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]203<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]204while being expressive, typed, and extensible at the same time. It
205aims to be a "universal IR" of sorts, by being at a low enough level
206that high-level ideas may be cleanly mapped to it (similar to how
207microprocessors are "universal IR's", allowing many source languages to
208be mapped to them). By providing type information, LLVM can be used as
209the target of optimizations: for example, through pointer analysis, it
210can be proven that a C automatic variable is never accessed outside of
211the current function... allowing it to be promoted to a simple SSA
212value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]213
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]214</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]215
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]216<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]217<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]218
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]219<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]220
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]221<p>It is important to note that this document describes 'well formed'
222LLVM assembly language. There is a difference between what the parser
223accepts and what is considered 'well formed'. For example, the
224following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]225
226<pre>
227 %x = <a href="#i_add">add</a> int 1, %x
228</pre>
229
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]230<p>...because the definition of <tt>%x</tt> does not dominate all of
231its uses. The LLVM infrastructure provides a verification pass that may
232be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]233automatically run by the parser after parsing input assembly and by
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]234the optimizer before it outputs bytecode. The violations pointed out
235by the verifier pass indicate bugs in transformation passes or input to
236the parser.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]237
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]238<!-- Describe the typesetting conventions here. --> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]239
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]240<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]241<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]242<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]243
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]244<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]245
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]246<p>LLVM uses three different forms of identifiers, for different
247purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]248
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]249<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]250 <li>Named values are represented as a string of characters with a '%' prefix.
251 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
252 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
253 Identifiers which require other characters in their names can be surrounded
254 with quotes. In this way, anything except a <tt>"</tt> character can be used
255 in a name.</li>
256
257 <li>Unnamed values are represented as an unsigned numeric value with a '%'
258 prefix. For example, %12, %2, %44.</li>
259
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]260 <li>Constants, which are described in a <a href="#constants">section about
261 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]262</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]263
264<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
265don't need to worry about name clashes with reserved words, and the set of
266reserved words may be expanded in the future without penalty. Additionally,
267unnamed identifiers allow a compiler to quickly come up with a temporary
268variable without having to avoid symbol table conflicts.</p>
269
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]270<p>Reserved words in LLVM are very similar to reserved words in other
271languages. There are keywords for different opcodes ('<tt><a
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]272href="#i_add">add</a></tt>', '<tt><a href="#i_cast">cast</a></tt>', '<tt><a
273href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
274href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
275and others. These reserved words cannot conflict with variable names, because
276none of them start with a '%' character.</p>
277
278<p>Here is an example of LLVM code to multiply the integer variable
279'<tt>%X</tt>' by 8:</p>
280
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]281<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]282
283<pre>
284 %result = <a href="#i_mul">mul</a> uint %X, 8
285</pre>
286
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]287<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]288
289<pre>
290 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
291</pre>
292
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]293<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]294
295<pre>
296 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
297 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
298 %result = <a href="#i_add">add</a> uint %1, %1
299</pre>
300
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]301<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
302important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]303
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]304<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]305
306 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
307 line.</li>
308
309 <li>Unnamed temporaries are created when the result of a computation is not
310 assigned to a named value.</li>
311
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]312 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]313
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]314</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]315
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]316<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]317demonstrating instructions, we will follow an instruction with a comment that
318defines the type and name of value produced. Comments are shown in italic
319text.</p>
320
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]321</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]322
323<!-- *********************************************************************** -->
324<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
325<!-- *********************************************************************** -->
326
327<!-- ======================================================================= -->
328<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
329</div>
330
331<div class="doc_text">
332
333<p>LLVM programs are composed of "Module"s, each of which is a
334translation unit of the input programs. Each module consists of
335functions, global variables, and symbol table entries. Modules may be
336combined together with the LLVM linker, which merges function (and
337global variable) definitions, resolves forward declarations, and merges
338symbol table entries. Here is an example of the "hello world" module:</p>
339
340<pre><i>; Declare the string constant as a global constant...</i>
341<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
342 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
343
344<i>; External declaration of the puts function</i>
345<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
346
347<i>; Definition of main function</i>
348int %main() { <i>; int()* </i>
349 <i>; Convert [13x sbyte]* to sbyte *...</i>
350 %cast210 = <a
351 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
352
353 <i>; Call puts function to write out the string to stdout...</i>
354 <a
355 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
356 <a
357 href="#i_ret">ret</a> int 0<br>}<br></pre>
358
359<p>This example is made up of a <a href="#globalvars">global variable</a>
360named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
361function, and a <a href="#functionstructure">function definition</a>
362for "<tt>main</tt>".</p>
363
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]364<p>In general, a module is made up of a list of global values,
365where both functions and global variables are global values. Global values are
366represented by a pointer to a memory location (in this case, a pointer to an
367array of char, and a pointer to a function), and have one of the following <a
368href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]369
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]370</div>
371
372<!-- ======================================================================= -->
373<div class="doc_subsection">
374 <a name="linkage">Linkage Types</a>
375</div>
376
377<div class="doc_text">
378
379<p>
380All Global Variables and Functions have one of the following types of linkage:
381</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]382
383<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]384
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]385 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]386
387 <dd>Global values with internal linkage are only directly accessible by
388 objects in the current module. In particular, linking code into a module with
389 an internal global value may cause the internal to be renamed as necessary to
390 avoid collisions. Because the symbol is internal to the module, all
391 references can be updated. This corresponds to the notion of the
392 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]393 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]394
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]395 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]396
397 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
398 the twist that linking together two modules defining the same
399 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
400 is typically used to implement inline functions. Unreferenced
401 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]402 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]403
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]404 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]405
406 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
407 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
408 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]409 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]410
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]411 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]412
413 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
414 pointer to array type. When two global variables with appending linkage are
415 linked together, the two global arrays are appended together. This is the
416 LLVM, typesafe, equivalent of having the system linker append together
417 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]418 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]419
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]420 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]421
422 <dd>If none of the above identifiers are used, the global is externally
423 visible, meaning that it participates in linkage and can be used to resolve
424 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]425 </dd>
426</dl>
427
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]428<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
429variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
430variable and was linked with this one, one of the two would be renamed,
431preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
432external (i.e., lacking any linkage declarations), they are accessible
433outside of the current module. It is illegal for a function <i>declaration</i>
434to have any linkage type other than "externally visible".</a></p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]435
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]436</div>
437
438<!-- ======================================================================= -->
439<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]440 <a name="callingconv">Calling Conventions</a>
441</div>
442
443<div class="doc_text">
444
445<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
446and <a href="#i_invoke">invokes</a> can all have an optional calling convention
447specified for the call. The calling convention of any pair of dynamic
448caller/callee must match, or the behavior of the program is undefined. The
449following calling conventions are supported by LLVM, and more may be added in
450the future:</p>
451
452<dl>
453 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
454
455 <dd>This calling convention (the default if no other calling convention is
456 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]457 supports varargs function calls and tolerates some mismatch in the declared
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]458 prototype and implemented declaration of the function (as does normal C).
459 </dd>
460
461 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
462
463 <dd>This calling convention attempts to make calls as fast as possible
464 (e.g. by passing things in registers). This calling convention allows the
465 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000[diff] [blame]466 without having to conform to an externally specified ABI. Implementations of
467 this convention should allow arbitrary tail call optimization to be supported.
468 This calling convention does not support varargs and requires the prototype of
469 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]470 </dd>
471
472 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
473
474 <dd>This calling convention attempts to make code in the caller as efficient
475 as possible under the assumption that the call is not commonly executed. As
476 such, these calls often preserve all registers so that the call does not break
477 any live ranges in the caller side. This calling convention does not support
478 varargs and requires the prototype of all callees to exactly match the
479 prototype of the function definition.
480 </dd>
481
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]482 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]483
484 <dd>Any calling convention may be specified by number, allowing
485 target-specific calling conventions to be used. Target specific calling
486 conventions start at 64.
487 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]488</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]489
490<p>More calling conventions can be added/defined on an as-needed basis, to
491support pascal conventions or any other well-known target-independent
492convention.</p>
493
494</div>
495
496<!-- ======================================================================= -->
497<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]498 <a name="globalvars">Global Variables</a>
499</div>
500
501<div class="doc_text">
502
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]503<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]504instead of run-time. Global variables may optionally be initialized, may have
505an explicit section to be placed in, and may
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]506have an optional explicit alignment specified. A
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]507variable may be defined as a global "constant," which indicates that the
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]508contents of the variable will <b>never</b> be modified (enabling better
509optimization, allowing the global data to be placed in the read-only section of
510an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]511cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]512
513<p>
514LLVM explicitly allows <em>declarations</em> of global variables to be marked
515constant, even if the final definition of the global is not. This capability
516can be used to enable slightly better optimization of the program, but requires
517the language definition to guarantee that optimizations based on the
518'constantness' are valid for the translation units that do not include the
519definition.
520</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]521
522<p>As SSA values, global variables define pointer values that are in
523scope (i.e. they dominate) all basic blocks in the program. Global
524variables always define a pointer to their "content" type because they
525describe a region of memory, and all memory objects in LLVM are
526accessed through pointers.</p>
527
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]528<p>LLVM allows an explicit section to be specified for globals. If the target
529supports it, it will emit globals to the section specified.</p>
530
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]531<p>An explicit alignment may be specified for a global. If not present, or if
532the alignment is set to zero, the alignment of the global is set by the target
533to whatever it feels convenient. If an explicit alignment is specified, the
534global is forced to have at least that much alignment. All alignments must be
535a power of 2.</p>
536
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]537</div>
538
539
540<!-- ======================================================================= -->
541<div class="doc_subsection">
542 <a name="functionstructure">Functions</a>
543</div>
544
545<div class="doc_text">
546
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]547<p>LLVM function definitions consist of an optional <a href="#linkage">linkage
548type</a>, an optional <a href="#callingconv">calling convention</a>, a return
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]549type, a function name, a (possibly empty) argument list, an optional section,
550an optional alignment, an opening curly brace,
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]551a list of basic blocks, and a closing curly brace. LLVM function declarations
552are defined with the "<tt>declare</tt>" keyword, an optional <a
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]553href="#callingconv">calling convention</a>, a return type, a function name,
554a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]555
556<p>A function definition contains a list of basic blocks, forming the CFG for
557the function. Each basic block may optionally start with a label (giving the
558basic block a symbol table entry), contains a list of instructions, and ends
559with a <a href="#terminators">terminator</a> instruction (such as a branch or
560function return).</p>
561
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]562<p>The first basic block in a program is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]563executed on entrance to the function, and it is not allowed to have predecessor
564basic blocks (i.e. there can not be any branches to the entry block of a
565function). Because the block can have no predecessors, it also cannot have any
566<a href="#i_phi">PHI nodes</a>.</p>
567
568<p>LLVM functions are identified by their name and type signature. Hence, two
569functions with the same name but different parameter lists or return values are
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]570considered different functions, and LLVM will resolve references to each
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]571appropriately.</p>
572
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]573<p>LLVM allows an explicit section to be specified for functions. If the target
574supports it, it will emit functions to the section specified.</p>
575
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]576<p>An explicit alignment may be specified for a function. If not present, or if
577the alignment is set to zero, the alignment of the function is set by the target
578to whatever it feels convenient. If an explicit alignment is specified, the
579function is forced to have at least that much alignment. All alignments must be
580a power of 2.</p>
581
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]582</div>
583
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame^]584<!-- ======================================================================= -->
585<div class="doc_subsection">
586 <a name="moduleasm">Module-Level Inline Assembly</a></li>
587</div>
588
589<div class="doc_text">
590<p>
591Modules may contain "module-level inline asm" blocks, which corresponds to the
592GCC "file scope inline asm" blocks. These blocks are internally concatenated by
593LLVM and treated as a single unit, but may be separated in the .ll file if
594desired. The syntax is very simple:
595</p>
596
597<div class="doc_code"><pre>
598 asm "inline asm code goes here"
599 asm "more can go here"
600</pre></div>
601
602<p>The strings can contain any character by escaping non-printable characters.
603 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
604 for the number.
605</p>
606
607<p>
608 The inline asm code is simply printed to the machine code .s file when
609 assembly code is generated.
610</p>
611</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]612
613
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]614<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]615<div class="doc_section"> <a name="typesystem">Type System</a> </div>
616<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]617
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]618<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]619
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]620<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]621intermediate representation. Being typed enables a number of
622optimizations to be performed on the IR directly, without having to do
623extra analyses on the side before the transformation. A strong type
624system makes it easier to read the generated code and enables novel
625analyses and transformations that are not feasible to perform on normal
626three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]627
628</div>
629
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]630<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]631<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]632<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]633<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]634system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]635
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]636<table class="layout">
637 <tr class="layout">
638 <td class="left">
639 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]640 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]641 <tr><th>Type</th><th>Description</th></tr>
642 <tr><td><tt>void</tt></td><td>No value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000[diff] [blame]643 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
644 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
645 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
646 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
647 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]648 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]649 </tbody>
650 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]651 </td>
652 <td class="right">
653 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]654 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]655 <tr><th>Type</th><th>Description</th></tr>
656 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000[diff] [blame]657 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
658 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
659 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
660 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
661 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]662 </tbody>
663 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]664 </td>
665 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]666</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]667</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]668
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]669<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]670<div class="doc_subsubsection"> <a name="t_classifications">Type
671Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]672<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]673<p>These different primitive types fall into a few useful
674classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]675
676<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]677 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]678 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]679 <tr>
680 <td><a name="t_signed">signed</a></td>
681 <td><tt>sbyte, short, int, long, float, double</tt></td>
682 </tr>
683 <tr>
684 <td><a name="t_unsigned">unsigned</a></td>
685 <td><tt>ubyte, ushort, uint, ulong</tt></td>
686 </tr>
687 <tr>
688 <td><a name="t_integer">integer</a></td>
689 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
690 </tr>
691 <tr>
692 <td><a name="t_integral">integral</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]693 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
694 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]695 </tr>
696 <tr>
697 <td><a name="t_floating">floating point</a></td>
698 <td><tt>float, double</tt></td>
699 </tr>
700 <tr>
701 <td><a name="t_firstclass">first class</a></td>
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]702 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
703 float, double, <a href="#t_pointer">pointer</a>,
704 <a href="#t_packed">packed</a></tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]705 </tr>
706 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]707</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]708
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]709<p>The <a href="#t_firstclass">first class</a> types are perhaps the
710most important. Values of these types are the only ones which can be
711produced by instructions, passed as arguments, or used as operands to
712instructions. This means that all structures and arrays must be
713manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]714</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]715
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]716<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]717<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]718
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]719<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]720
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]721<p>The real power in LLVM comes from the derived types in the system.
722This is what allows a programmer to represent arrays, functions,
723pointers, and other useful types. Note that these derived types may be
724recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]725
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]726</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]727
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]728<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]729<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]730
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]731<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]732
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]733<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]734
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]735<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]736sequentially in memory. The array type requires a size (number of
737elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]738
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]739<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]740
741<pre>
742 [&lt;# elements&gt; x &lt;elementtype&gt;]
743</pre>
744
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]745<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]746be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]747
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]748<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]749<table class="layout">
750 <tr class="layout">
751 <td class="left">
752 <tt>[40 x int ]</tt><br/>
753 <tt>[41 x int ]</tt><br/>
754 <tt>[40 x uint]</tt><br/>
755 </td>
756 <td class="left">
757 Array of 40 integer values.<br/>
758 Array of 41 integer values.<br/>
759 Array of 40 unsigned integer values.<br/>
760 </td>
761 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]762</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]763<p>Here are some examples of multidimensional arrays:</p>
764<table class="layout">
765 <tr class="layout">
766 <td class="left">
767 <tt>[3 x [4 x int]]</tt><br/>
768 <tt>[12 x [10 x float]]</tt><br/>
769 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
770 </td>
771 <td class="left">
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]772 3x4 array of integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]773 12x10 array of single precision floating point values.<br/>
774 2x3x4 array of unsigned integer values.<br/>
775 </td>
776 </tr>
777</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]778
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]779<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
780length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]781LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
782As a special case, however, zero length arrays are recognized to be variable
783length. This allows implementation of 'pascal style arrays' with the LLVM
784type "{ int, [0 x float]}", for example.</p>
785
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]786</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]787
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]788<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]789<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]790<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]791<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]792<p>The function type can be thought of as a function signature. It
793consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]794Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]795(which are structures of pointers to functions), for indirect function
796calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]797<p>
798The return type of a function type cannot be an aggregate type.
799</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]800<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]801<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]802<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]803specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]804which indicates that the function takes a variable number of arguments.
805Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]806 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]807<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]808<table class="layout">
809 <tr class="layout">
810 <td class="left">
811 <tt>int (int)</tt> <br/>
812 <tt>float (int, int *) *</tt><br/>
813 <tt>int (sbyte *, ...)</tt><br/>
814 </td>
815 <td class="left">
816 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
817 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]818 <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]819 returning <tt>float</tt>.<br/>
820 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
821 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
822 the signature for <tt>printf</tt> in LLVM.<br/>
823 </td>
824 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]825</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]826
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]827</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]828<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]829<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]830<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]831<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]832<p>The structure type is used to represent a collection of data members
833together in memory. The packing of the field types is defined to match
834the ABI of the underlying processor. The elements of a structure may
835be any type that has a size.</p>
836<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
837and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
838field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
839instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]840<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]841<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]842<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]843<table class="layout">
844 <tr class="layout">
845 <td class="left">
846 <tt>{ int, int, int }</tt><br/>
847 <tt>{ float, int (int) * }</tt><br/>
848 </td>
849 <td class="left">
850 a triple of three <tt>int</tt> values<br/>
851 A pair, where the first element is a <tt>float</tt> and the second element
852 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
853 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
854 </td>
855 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]856</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]857</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]858
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]859<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]860<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]861<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]862<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]863<p>As in many languages, the pointer type represents a pointer or
864reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]865<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]866<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]867<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]868<table class="layout">
869 <tr class="layout">
870 <td class="left">
871 <tt>[4x int]*</tt><br/>
872 <tt>int (int *) *</tt><br/>
873 </td>
874 <td class="left">
875 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
876 four <tt>int</tt> values<br/>
877 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattnera977c482005-02-19 02:22:14 +0000[diff] [blame]878 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]879 <tt>int</tt>.<br/>
880 </td>
881 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]882</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]883</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]884
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]885<!-- _______________________________________________________________________ -->
886<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]887<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]888
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]889<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]890
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]891<p>A packed type is a simple derived type that represents a vector
892of elements. Packed types are used when multiple primitive data
893are operated in parallel using a single instruction (SIMD).
894A packed type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000[diff] [blame]895elements) and an underlying primitive data type. Vectors must have a power
896of two length (1, 2, 4, 8, 16 ...). Packed types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]897considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]898
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]899<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]900
901<pre>
902 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
903</pre>
904
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]905<p>The number of elements is a constant integer value; elementtype may
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]906be any integral or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]907
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]908<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]909
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]910<table class="layout">
911 <tr class="layout">
912 <td class="left">
913 <tt>&lt;4 x int&gt;</tt><br/>
914 <tt>&lt;8 x float&gt;</tt><br/>
915 <tt>&lt;2 x uint&gt;</tt><br/>
916 </td>
917 <td class="left">
918 Packed vector of 4 integer values.<br/>
919 Packed vector of 8 floating-point values.<br/>
920 Packed vector of 2 unsigned integer values.<br/>
921 </td>
922 </tr>
923</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]924</div>
925
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]926<!-- _______________________________________________________________________ -->
927<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
928<div class="doc_text">
929
930<h5>Overview:</h5>
931
932<p>Opaque types are used to represent unknown types in the system. This
933corresponds (for example) to the C notion of a foward declared structure type.
934In LLVM, opaque types can eventually be resolved to any type (not just a
935structure type).</p>
936
937<h5>Syntax:</h5>
938
939<pre>
940 opaque
941</pre>
942
943<h5>Examples:</h5>
944
945<table class="layout">
946 <tr class="layout">
947 <td class="left">
948 <tt>opaque</tt>
949 </td>
950 <td class="left">
951 An opaque type.<br/>
952 </td>
953 </tr>
954</table>
955</div>
956
957
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]958<!-- *********************************************************************** -->
959<div class="doc_section"> <a name="constants">Constants</a> </div>
960<!-- *********************************************************************** -->
961
962<div class="doc_text">
963
964<p>LLVM has several different basic types of constants. This section describes
965them all and their syntax.</p>
966
967</div>
968
969<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]970<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]971
972<div class="doc_text">
973
974<dl>
975 <dt><b>Boolean constants</b></dt>
976
977 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
978 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
979 </dd>
980
981 <dt><b>Integer constants</b></dt>
982
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]983 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]984 href="#t_integer">integer</a> type. Negative numbers may be used with signed
985 integer types.
986 </dd>
987
988 <dt><b>Floating point constants</b></dt>
989
990 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
991 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]992 notation (see below). Floating point constants must have a <a
993 href="#t_floating">floating point</a> type. </dd>
994
995 <dt><b>Null pointer constants</b></dt>
996
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]997 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]998 and must be of <a href="#t_pointer">pointer type</a>.</dd>
999
1000</dl>
1001
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1002<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1003of floating point constants. For example, the form '<tt>double
10040x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
10054.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]1006(and the only time that they are generated by the disassembler) is when a
1007floating point constant must be emitted but it cannot be represented as a
1008decimal floating point number. For example, NaN's, infinities, and other
1009special values are represented in their IEEE hexadecimal format so that
1010assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1011
1012</div>
1013
1014<!-- ======================================================================= -->
1015<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1016</div>
1017
1018<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]1019<p>Aggregate constants arise from aggregation of simple constants
1020and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1021
1022<dl>
1023 <dt><b>Structure constants</b></dt>
1024
1025 <dd>Structure constants are represented with notation similar to structure
1026 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]1027 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
1028 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
1029 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1030 types of elements must match those specified by the type.
1031 </dd>
1032
1033 <dt><b>Array constants</b></dt>
1034
1035 <dd>Array constants are represented with notation similar to array type
1036 definitions (a comma separated list of elements, surrounded by square brackets
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1037 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1038 constants must have <a href="#t_array">array type</a>, and the number and
1039 types of elements must match those specified by the type.
1040 </dd>
1041
1042 <dt><b>Packed constants</b></dt>
1043
1044 <dd>Packed constants are represented with notation similar to packed type
1045 definitions (a comma separated list of elements, surrounded by
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1046 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]1047 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
1048 href="#t_packed">packed type</a>, and the number and types of elements must
1049 match those specified by the type.
1050 </dd>
1051
1052 <dt><b>Zero initialization</b></dt>
1053
1054 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1055 value to zero of <em>any</em> type, including scalar and aggregate types.
1056 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1057 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1058 initializers.
1059 </dd>
1060</dl>
1061
1062</div>
1063
1064<!-- ======================================================================= -->
1065<div class="doc_subsection">
1066 <a name="globalconstants">Global Variable and Function Addresses</a>
1067</div>
1068
1069<div class="doc_text">
1070
1071<p>The addresses of <a href="#globalvars">global variables</a> and <a
1072href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1073constants. These constants are explicitly referenced when the <a
1074href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1075href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1076file:</p>
1077
1078<pre>
1079 %X = global int 17
1080 %Y = global int 42
1081 %Z = global [2 x int*] [ int* %X, int* %Y ]
1082</pre>
1083
1084</div>
1085
1086<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1087<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1088<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1089 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1090 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1091 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1092
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1093 <p>Undefined values indicate to the compiler that the program is well defined
1094 no matter what value is used, giving the compiler more freedom to optimize.
1095 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1096</div>
1097
1098<!-- ======================================================================= -->
1099<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1100</div>
1101
1102<div class="doc_text">
1103
1104<p>Constant expressions are used to allow expressions involving other constants
1105to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1106href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1107that does not have side effects (e.g. load and call are not supported). The
1108following is the syntax for constant expressions:</p>
1109
1110<dl>
1111 <dt><b><tt>cast ( CST to TYPE )</tt></b></dt>
1112
1113 <dd>Cast a constant to another type.</dd>
1114
1115 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1116
1117 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1118 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1119 instruction, the index list may have zero or more indexes, which are required
1120 to make sense for the type of "CSTPTR".</dd>
1121
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]1122 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1123
1124 <dd>Perform the <a href="#i_select">select operation</a> on
1125 constants.
1126
1127 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1128
1129 <dd>Perform the <a href="#i_extractelement">extractelement
1130 operation</a> on constants.
1131
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]1132 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1133
1134 <dd>Perform the <a href="#i_insertelement">insertelement
1135 operation</a> on constants.
1136
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1137 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1138
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1139 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1140 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1141 binary</a> operations. The constraints on operands are the same as those for
1142 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1143 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1144</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1145</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]1146
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1147<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1148<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1149<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1150
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1151<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1152
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1153<p>The LLVM instruction set consists of several different
1154classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1155instructions</a>, <a href="#binaryops">binary instructions</a>,
1156<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1157 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1158instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1159
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1160</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1161
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1162<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1163<div class="doc_subsection"> <a name="terminators">Terminator
1164Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1165
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1166<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1167
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1168<p>As mentioned <a href="#functionstructure">previously</a>, every
1169basic block in a program ends with a "Terminator" instruction, which
1170indicates which block should be executed after the current block is
1171finished. These terminator instructions typically yield a '<tt>void</tt>'
1172value: they produce control flow, not values (the one exception being
1173the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1174<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1175 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1176instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1177the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1178 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1179 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1180
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1181</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1182
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1183<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1184<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1185Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1186<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1187<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1188<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]1189 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1190</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1191<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1192<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1193value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]1194<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1195returns a value and then causes control flow, and one that just causes
1196control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1197<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1198<p>The '<tt>ret</tt>' instruction may return any '<a
1199 href="#t_firstclass">first class</a>' type. Notice that a function is
1200not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1201instruction inside of the function that returns a value that does not
1202match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1203<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1204<p>When the '<tt>ret</tt>' instruction is executed, control flow
1205returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +0000[diff] [blame]1206 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1207the instruction after the call. If the caller was an "<a
1208 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1209at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1210returns a value, that value shall set the call or invoke instruction's
1211return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1212<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1213<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1214 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1215</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1216</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1217<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1218<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1219<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1220<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1221<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]1222</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1223<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1224<p>The '<tt>br</tt>' instruction is used to cause control flow to
1225transfer to a different basic block in the current function. There are
1226two forms of this instruction, corresponding to a conditional branch
1227and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1228<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1229<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1230single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1231unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1232value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1233<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1234<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1235argument is evaluated. If the value is <tt>true</tt>, control flows
1236to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1237control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1238<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1239<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
1240 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]1241</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1242<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1243<div class="doc_subsubsection">
1244 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1245</div>
1246
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1247<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1248<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1249
1250<pre>
1251 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1252</pre>
1253
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1254<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1255
1256<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1257several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1258instruction, allowing a branch to occur to one of many possible
1259destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1260
1261
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1262<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1263
1264<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1265comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1266an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1267table is not allowed to contain duplicate constant entries.</p>
1268
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1269<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1270
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1271<p>The <tt>switch</tt> instruction specifies a table of values and
1272destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1273table is searched for the given value. If the value is found, control flow is
1274transfered to the corresponding destination; otherwise, control flow is
1275transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1276
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1277<h5>Implementation:</h5>
1278
1279<p>Depending on properties of the target machine and the particular
1280<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1281ways. For example, it could be generated as a series of chained conditional
1282branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1283
1284<h5>Example:</h5>
1285
1286<pre>
1287 <i>; Emulate a conditional br instruction</i>
1288 %Val = <a href="#i_cast">cast</a> bool %value to int
1289 switch int %Val, label %truedest [int 0, label %falsedest ]
1290
1291 <i>; Emulate an unconditional br instruction</i>
1292 switch uint 0, label %dest [ ]
1293
1294 <i>; Implement a jump table:</i>
1295 switch uint %val, label %otherwise [ uint 0, label %onzero
1296 uint 1, label %onone
1297 uint 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1298</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1299</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1300
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1301<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1302<div class="doc_subsubsection">
1303 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1304</div>
1305
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1306<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1307
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1308<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1309
1310<pre>
1311 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)
1312 to label &lt;normal label&gt; except label &lt;exception label&gt;
1313</pre>
1314
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1315<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1316
1317<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1318function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1319'<tt>normal</tt>' label or the
1320'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1321"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1322"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1323href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1324continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1325
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1326<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1327
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1328<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1329
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1330<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1331 <li>
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1332 The optional "cconv" marker indicates which <a href="callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1333 convention</a> the call should use. If none is specified, the call defaults
1334 to using C calling conventions.
1335 </li>
1336 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1337 function value being invoked. In most cases, this is a direct function
1338 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1339 an arbitrary pointer to function value.
1340 </li>
1341
1342 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1343 function to be invoked. </li>
1344
1345 <li>'<tt>function args</tt>': argument list whose types match the function
1346 signature argument types. If the function signature indicates the function
1347 accepts a variable number of arguments, the extra arguments can be
1348 specified. </li>
1349
1350 <li>'<tt>normal label</tt>': the label reached when the called function
1351 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1352
1353 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1354 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1355
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1356</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1357
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1358<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1359
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1360<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1361href="#i_call">call</a></tt>' instruction in most regards. The primary
1362difference is that it establishes an association with a label, which is used by
1363the runtime library to unwind the stack.</p>
1364
1365<p>This instruction is used in languages with destructors to ensure that proper
1366cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1367exception. Additionally, this is important for implementation of
1368'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1369
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1370<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1371<pre>
1372 %retval = invoke int %Test(int 15) to label %Continue
1373 except label %TestCleanup <i>; {int}:retval set</i>
1374 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
1375 except label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1376</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1377</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1378
1379
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1380<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1381
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1382<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1383Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1384
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1385<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1386
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1387<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1388<pre>
1389 unwind
1390</pre>
1391
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1392<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1393
1394<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1395at the first callee in the dynamic call stack which used an <a
1396href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1397primarily used to implement exception handling.</p>
1398
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1399<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1400
1401<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1402immediately halt. The dynamic call stack is then searched for the first <a
1403href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1404execution continues at the "exceptional" destination block specified by the
1405<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1406dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1407</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1408
1409<!-- _______________________________________________________________________ -->
1410
1411<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1412Instruction</a> </div>
1413
1414<div class="doc_text">
1415
1416<h5>Syntax:</h5>
1417<pre>
1418 unreachable
1419</pre>
1420
1421<h5>Overview:</h5>
1422
1423<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1424instruction is used to inform the optimizer that a particular portion of the
1425code is not reachable. This can be used to indicate that the code after a
1426no-return function cannot be reached, and other facts.</p>
1427
1428<h5>Semantics:</h5>
1429
1430<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1431</div>
1432
1433
1434
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1435<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1436<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1437<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1438<p>Binary operators are used to do most of the computation in a
1439program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1440produce a single value. The operands might represent
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1441multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1442The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1443necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1444<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1445</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1446<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1447<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1448Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1449<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1450<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1451<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]1452</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1453<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1454<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1455<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1456<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1457 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1458 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1459Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1460<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1461<p>The value produced is the integer or floating point sum of the two
1462operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1463<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1464<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]1465</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1466</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1467<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1468<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1469Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1470<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1471<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1472<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]1473</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1474<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1475<p>The '<tt>sub</tt>' instruction returns the difference of its two
1476operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1477<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1478instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1479<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1480<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1481 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1482values.
1483This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1484Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1485<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1486<p>The value produced is the integer or floating point difference of
1487the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1488<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1489<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]1490 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1491</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1492</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1493<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1494<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1495Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1496<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1497<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1498<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]1499</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1500<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1501<p>The '<tt>mul</tt>' instruction returns the product of its two
1502operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1503<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1504<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1505 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1506values.
1507This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1508Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1509<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1510<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1511two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1512<p>There is no signed vs unsigned multiplication. The appropriate
1513action is taken based on the type of the operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1514<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1515<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]1516</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1517</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1518<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1519<div class="doc_subsubsection"> <a name="i_div">'<tt>div</tt>'
1520Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1521<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1522<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1523<pre> &lt;result&gt; = div &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1524</pre>
1525<h5>Overview:</h5>
1526<p>The '<tt>div</tt>' instruction returns the quotient of its two
1527operands.</p>
1528<h5>Arguments:</h5>
1529<p>The two arguments to the '<tt>div</tt>' instruction must be either <a
1530 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1531values.
1532This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1533Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1534<h5>Semantics:</h5>
1535<p>The value produced is the integer or floating point quotient of the
1536two operands.</p>
1537<h5>Example:</h5>
1538<pre> &lt;result&gt; = div int 4, %var <i>; yields {int}:result = 4 / %var</i>
1539</pre>
1540</div>
1541<!-- _______________________________________________________________________ -->
1542<div class="doc_subsubsection"> <a name="i_rem">'<tt>rem</tt>'
1543Instruction</a> </div>
1544<div class="doc_text">
1545<h5>Syntax:</h5>
1546<pre> &lt;result&gt; = rem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1547</pre>
1548<h5>Overview:</h5>
1549<p>The '<tt>rem</tt>' instruction returns the remainder from the
1550division of its two operands.</p>
1551<h5>Arguments:</h5>
1552<p>The two arguments to the '<tt>rem</tt>' instruction must be either <a
1553 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1554values.
1555This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1556Both arguments must have identical types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1557<h5>Semantics:</h5>
1558<p>This returns the <i>remainder</i> of a division (where the result
1559has the same sign as the divisor), not the <i>modulus</i> (where the
1560result has the same sign as the dividend) of a value. For more
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1561information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1562 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1563Math Forum</a>.</p>
1564<h5>Example:</h5>
1565<pre> &lt;result&gt; = rem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1566</pre>
1567</div>
1568<!-- _______________________________________________________________________ -->
1569<div class="doc_subsubsection"> <a name="i_setcc">'<tt>set<i>cc</i></tt>'
1570Instructions</a> </div>
1571<div class="doc_text">
1572<h5>Syntax:</h5>
1573<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]1574 &lt;result&gt; = setne &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1575 &lt;result&gt; = setlt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1576 &lt;result&gt; = setgt &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1577 &lt;result&gt; = setle &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1578 &lt;result&gt; = setge &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
1579</pre>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1580<h5>Overview:</h5>
1581<p>The '<tt>set<i>cc</i></tt>' family of instructions returns a boolean
1582value based on a comparison of their two operands.</p>
1583<h5>Arguments:</h5>
1584<p>The two arguments to the '<tt>set<i>cc</i></tt>' instructions must
1585be of <a href="#t_firstclass">first class</a> type (it is not possible
1586to compare '<tt>label</tt>'s, '<tt>array</tt>'s, '<tt>structure</tt>'
1587or '<tt>void</tt>' values, etc...). Both arguments must have identical
1588types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1589<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1590<p>The '<tt>seteq</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1591value if both operands are equal.<br>
1592The '<tt>setne</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1593value if both operands are unequal.<br>
1594The '<tt>setlt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1595value if the first operand is less than the second operand.<br>
1596The '<tt>setgt</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1597value if the first operand is greater than the second operand.<br>
1598The '<tt>setle</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1599value if the first operand is less than or equal to the second operand.<br>
1600The '<tt>setge</tt>' instruction yields a <tt>true</tt> '<tt>bool</tt>'
1601value if the first operand is greater than or equal to the second
1602operand.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1603<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1604<pre> &lt;result&gt; = seteq int 4, 5 <i>; yields {bool}:result = false</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1605 &lt;result&gt; = setne float 4, 5 <i>; yields {bool}:result = true</i>
1606 &lt;result&gt; = setlt uint 4, 5 <i>; yields {bool}:result = true</i>
1607 &lt;result&gt; = setgt sbyte 4, 5 <i>; yields {bool}:result = false</i>
1608 &lt;result&gt; = setle sbyte 4, 5 <i>; yields {bool}:result = true</i>
1609 &lt;result&gt; = setge sbyte 4, 5 <i>; yields {bool}:result = false</i>
1610</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1611</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1612<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1613<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1614Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1615<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1616<p>Bitwise binary operators are used to do various forms of
1617bit-twiddling in a program. They are generally very efficient
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1618instructions and can commonly be strength reduced from other
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1619instructions. They require two operands, execute an operation on them,
1620and produce a single value. The resulting value of the bitwise binary
1621operators is always the same type as its first operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1622</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1623<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1624<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1625Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1626<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1627<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1628<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]1629</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1630<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1631<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1632its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1633<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1634<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1635 href="#t_integral">integral</a> values. Both arguments must have
1636identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1637<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1638<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1639<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1640<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1641<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1642 <tbody>
1643 <tr>
1644 <td>In0</td>
1645 <td>In1</td>
1646 <td>Out</td>
1647 </tr>
1648 <tr>
1649 <td>0</td>
1650 <td>0</td>
1651 <td>0</td>
1652 </tr>
1653 <tr>
1654 <td>0</td>
1655 <td>1</td>
1656 <td>0</td>
1657 </tr>
1658 <tr>
1659 <td>1</td>
1660 <td>0</td>
1661 <td>0</td>
1662 </tr>
1663 <tr>
1664 <td>1</td>
1665 <td>1</td>
1666 <td>1</td>
1667 </tr>
1668 </tbody>
1669</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1670</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1671<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1672<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]1673 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1674 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1675</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1676</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1677<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1678<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1679<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1680<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1681<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]1682</pre>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1683<h5>Overview:</h5>
1684<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1685or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1686<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1687<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1688 href="#t_integral">integral</a> values. Both arguments must have
1689identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1690<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1691<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1692<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1693<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1694<table border="1" cellspacing="0" cellpadding="4">
1695 <tbody>
1696 <tr>
1697 <td>In0</td>
1698 <td>In1</td>
1699 <td>Out</td>
1700 </tr>
1701 <tr>
1702 <td>0</td>
1703 <td>0</td>
1704 <td>0</td>
1705 </tr>
1706 <tr>
1707 <td>0</td>
1708 <td>1</td>
1709 <td>1</td>
1710 </tr>
1711 <tr>
1712 <td>1</td>
1713 <td>0</td>
1714 <td>1</td>
1715 </tr>
1716 <tr>
1717 <td>1</td>
1718 <td>1</td>
1719 <td>1</td>
1720 </tr>
1721 </tbody>
1722</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1723</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1724<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1725<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]1726 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1727 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1728</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1729</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1730<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1731<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
1732Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1733<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1734<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1735<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]1736</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1737<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1738<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
1739or of its two operands. The <tt>xor</tt> is used to implement the
1740"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1741<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1742<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1743 href="#t_integral">integral</a> values. Both arguments must have
1744identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1745<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1746<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1747<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1748<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1749<table border="1" cellspacing="0" cellpadding="4">
1750 <tbody>
1751 <tr>
1752 <td>In0</td>
1753 <td>In1</td>
1754 <td>Out</td>
1755 </tr>
1756 <tr>
1757 <td>0</td>
1758 <td>0</td>
1759 <td>0</td>
1760 </tr>
1761 <tr>
1762 <td>0</td>
1763 <td>1</td>
1764 <td>1</td>
1765 </tr>
1766 <tr>
1767 <td>1</td>
1768 <td>0</td>
1769 <td>1</td>
1770 </tr>
1771 <tr>
1772 <td>1</td>
1773 <td>1</td>
1774 <td>0</td>
1775 </tr>
1776 </tbody>
1777</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1778</div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1779<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1780<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1781<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]1782 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
1783 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1784 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1785</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1786</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1787<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1788<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
1789Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1790<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1791<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1792<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]1793</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1794<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1795<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
1796the left a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1797<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1798<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1799 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1800type.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1801<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1802<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]1803<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1804<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]1805 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
1806 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
1807</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1808</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1809<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1810<div class="doc_subsubsection"> <a name="i_shr">'<tt>shr</tt>'
1811Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1812<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1813<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1814<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]1815</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1816<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1817<p>The '<tt>shr</tt>' instruction returns the first operand shifted to
1818the right a specified number of bits.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1819<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1820<p>The first argument to the '<tt>shr</tt>' instruction must be an <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1821 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
1822type.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1823<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1824<p>If the first argument is a <a href="#t_signed">signed</a> type, the
1825most significant bit is duplicated in the newly free'd bit positions.
1826If the first argument is unsigned, zero bits shall fill the empty
1827positions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1828<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1829<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]1830 &lt;result&gt; = shr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1831 &lt;result&gt; = shr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
Chris Lattner8c6bb902003-06-18 21:30:51 +0000[diff] [blame]1832 &lt;result&gt; = shr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
1833 &lt;result&gt; = shr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1834</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1835</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1836
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1837<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1838<div class="doc_subsection">
1839 <a name="memoryops">Memory Access Operations</a>
1840</div>
1841
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1842<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1843
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1844<p>A key design point of an SSA-based representation is how it
1845represents memory. In LLVM, no memory locations are in SSA form, which
1846makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1847allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1848
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1849</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1850
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1851<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1852<div class="doc_subsubsection">
1853 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
1854</div>
1855
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1856<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1857
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1858<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1859
1860<pre>
1861 &lt;result&gt; = malloc &lt;type&gt;[, uint &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1862</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1863
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1864<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1865
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1866<p>The '<tt>malloc</tt>' instruction allocates memory from the system
1867heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1868
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1869<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1870
1871<p>The '<tt>malloc</tt>' instruction allocates
1872<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +0000[diff] [blame]1873bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1874appropriate type to the program. If "NumElements" is specified, it is the
1875number of elements allocated. If an alignment is specified, the value result
1876of the allocation is guaranteed to be aligned to at least that boundary. If
1877not specified, or if zero, the target can choose to align the allocation on any
1878convenient boundary.</p>
1879
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1880<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1881
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1882<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1883
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1884<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
1885a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1886
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1887<h5>Example:</h5>
1888
1889<pre>
1890 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
1891
1892 %size = <a 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]1893 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
1894 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1895 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
1896 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1897</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1898</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1899
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1900<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1901<div class="doc_subsubsection">
1902 <a name="i_free">'<tt>free</tt>' Instruction</a>
1903</div>
1904
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1905<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1906
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1907<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1908
1909<pre>
1910 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1911</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1912
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1913<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1914
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1915<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1916memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1917
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1918<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1919
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1920<p>'<tt>value</tt>' shall be a pointer value that points to a value
1921that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
1922instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1923
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1924<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1925
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1926<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1927after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1928
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1929<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1930
1931<pre>
1932 %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]1933 free [4 x ubyte]* %array
1934</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1935</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1936
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1937<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1938<div class="doc_subsubsection">
1939 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
1940</div>
1941
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1942<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1943
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1944<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1945
1946<pre>
1947 &lt;result&gt; = alloca &lt;type&gt;[, uint &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1948</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1949
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1950<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1951
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1952<p>The '<tt>alloca</tt>' instruction allocates memory on the current
1953stack frame of the procedure that is live until the current function
1954returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1955
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1956<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1957
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1958<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1959bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1960appropriate type to the program. If "NumElements" is specified, it is the
1961number of elements allocated. If an alignment is specified, the value result
1962of the allocation is guaranteed to be aligned to at least that boundary. If
1963not specified, or if zero, the target can choose to align the allocation on any
1964convenient boundary.</p>
1965
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1966<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1967
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1968<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1969
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1970<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1971memory is automatically released when the function returns. The '<tt>alloca</tt>'
1972instruction is commonly used to represent automatic variables that must
1973have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +0000[diff] [blame]1974 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1975instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1976
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1977<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1978
1979<pre>
1980 %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1981 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1982 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
1983 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1984</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1985</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]1986
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1987<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1988<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
1989Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1990<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1991<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1992<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]1993<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1994<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]1995<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1996<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1997address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +0000[diff] [blame]1998 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1999marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2000the number or order of execution of this <tt>load</tt> with other
2001volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2002instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2003<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2004<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2005<h5>Examples:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2006<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2007 <a
2008 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2009 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2010</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2011</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2012<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2013<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2014Instruction</a> </div>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2015<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2016<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]2017 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]2018</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2019<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2020<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2021<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2022<p>There are two arguments to the '<tt>store</tt>' instruction: a value
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2023to store and an address in which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2024operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2025operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2026optimizer is not allowed to modify the number or order of execution of
2027this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2028 href="#i_store">store</a></tt> instructions.</p>
2029<h5>Semantics:</h5>
2030<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2031at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2032<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2033<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2034 <a
2035 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2036 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2037</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2038<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2039<div class="doc_subsubsection">
2040 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2041</div>
2042
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2043<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2044<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2045<pre>
2046 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2047</pre>
2048
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2049<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2050
2051<p>
2052The '<tt>getelementptr</tt>' instruction is used to get the address of a
2053subelement of an aggregate data structure.</p>
2054
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2055<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2056
2057<p>This instruction takes a list of integer constants that indicate what
2058elements of the aggregate object to index to. The actual types of the arguments
2059provided depend on the type of the first pointer argument. The
2060'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2061levels of a structure or to a specific index in an array. When indexing into a
2062structure, only <tt>uint</tt>
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2063integer constants are allowed. When indexing into an array or pointer,
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2064<tt>int</tt> and <tt>long</tt> indexes are allowed of any sign.</p>
2065
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2066<p>For example, let's consider a C code fragment and how it gets
2067compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2068
2069<pre>
2070 struct RT {
2071 char A;
2072 int B[10][20];
2073 char C;
2074 };
2075 struct ST {
2076 int X;
2077 double Y;
2078 struct RT Z;
2079 };
2080
2081 int *foo(struct ST *s) {
2082 return &amp;s[1].Z.B[5][13];
2083 }
2084</pre>
2085
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2086<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2087
2088<pre>
2089 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2090 %ST = type { int, double, %RT }
2091
Brian Gaeke7283e7c2004-07-02 21:08:14 +0000[diff] [blame]2092 implementation
2093
2094 int* %foo(%ST* %s) {
2095 entry:
2096 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2097 ret int* %reg
2098 }
2099</pre>
2100
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2101<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2102
2103<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2104on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Chris Lattnere53e5082004-06-03 22:57:15 +0000[diff] [blame]2105and <a href="#t_array">array</a> types require <tt>uint</tt>, <tt>int</tt>,
2106<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]2107types require <tt>uint</tt> <b>constants</b>.</p>
2108
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2109<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]2110type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2111}</tt>' type, a structure. The second index indexes into the third element of
2112the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2113sbyte }</tt>' type, another structure. The third index indexes into the second
2114element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2115array. The two dimensions of the array are subscripted into, yielding an
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2116'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2117to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2118
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2119<p>Note that it is perfectly legal to index partially through a
2120structure, returning a pointer to an inner element. Because of this,
2121the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2122
2123<pre>
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]2124 int* %foo(%ST* %s) {
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2125 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2126 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2127 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2128 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2129 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2130 ret int* %t5
2131 }
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2132</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2133
2134<p>Note that it is undefined to access an array out of bounds: array and
2135pointer indexes must always be within the defined bounds of the array type.
2136The one exception for this rules is zero length arrays. These arrays are
2137defined to be accessible as variable length arrays, which requires access
2138beyond the zero'th element.</p>
2139
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2140<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2141
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2142<pre>
2143 <i>; yields [12 x ubyte]*:aptr</i>
2144 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2145</pre>
2146
2147</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2148<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2149<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2150<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]2151<p>The instructions in this category are the "miscellaneous"
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2152instructions, which defy better classification.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2153</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2154<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2155<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
2156Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2157<div class="doc_text">
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]2158<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2159<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]2160<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2161<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
2162the SSA graph representing the function.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]2163<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2164<p>The type of the incoming values are specified with the first type
2165field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
2166as arguments, with one pair for each predecessor basic block of the
2167current block. Only values of <a href="#t_firstclass">first class</a>
2168type may be used as the value arguments to the PHI node. Only labels
2169may be used as the label arguments.</p>
2170<p>There must be no non-phi instructions between the start of a basic
2171block and the PHI instructions: i.e. PHI instructions must be first in
2172a basic block.</p>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]2173<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2174<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
2175value specified by the parameter, depending on which basic block we
2176came from in the last <a href="#terminators">terminator</a> instruction.</p>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2177<h5>Example:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2178<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]2179</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2180
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2181<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2182<div class="doc_subsubsection">
2183 <a name="i_cast">'<tt>cast .. to</tt>' Instruction</a>
2184</div>
2185
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2186<div class="doc_text">
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2187
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2188<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2189
2190<pre>
2191 &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]2192</pre>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2193
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2194<h5>Overview:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2195
2196<p>
2197The '<tt>cast</tt>' instruction is used as the primitive means to convert
2198integers to floating point, change data type sizes, and break type safety (by
2199casting pointers).
2200</p>
2201
2202
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2203<h5>Arguments:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2204
2205<p>
2206The '<tt>cast</tt>' instruction takes a value to cast, which must be a first
2207class value, and a type to cast it to, which must also be a <a
2208href="#t_firstclass">first class</a> type.
2209</p>
2210
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2211<h5>Semantics:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2212
2213<p>
2214This instruction follows the C rules for explicit casts when determining how the
2215data being cast must change to fit in its new container.
2216</p>
2217
2218<p>
2219When casting to bool, any value that would be considered true in the context of
2220a C '<tt>if</tt>' condition is converted to the boolean '<tt>true</tt>' values,
2221all else are '<tt>false</tt>'.
2222</p>
2223
2224<p>
2225When extending an integral value from a type of one signness to another (for
2226example '<tt>sbyte</tt>' to '<tt>ulong</tt>'), the value is sign-extended if the
2227<b>source</b> value is signed, and zero-extended if the source value is
2228unsigned. <tt>bool</tt> values are always zero extended into either zero or
2229one.
2230</p>
2231
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]2232<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2233
2234<pre>
2235 %X = cast int 257 to ubyte <i>; yields ubyte:1</i>
Chris Lattner7bae3952002-06-25 18:03:17 +0000[diff] [blame]2236 %Y = cast int 123 to bool <i>; yields bool:true</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]2237</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2238</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2239
2240<!-- _______________________________________________________________________ -->
2241<div class="doc_subsubsection">
2242 <a name="i_select">'<tt>select</tt>' Instruction</a>
2243</div>
2244
2245<div class="doc_text">
2246
2247<h5>Syntax:</h5>
2248
2249<pre>
2250 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
2251</pre>
2252
2253<h5>Overview:</h5>
2254
2255<p>
2256The '<tt>select</tt>' instruction is used to choose one value based on a
2257condition, without branching.
2258</p>
2259
2260
2261<h5>Arguments:</h5>
2262
2263<p>
2264The '<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.
2265</p>
2266
2267<h5>Semantics:</h5>
2268
2269<p>
2270If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2271value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2272</p>
2273
2274<h5>Example:</h5>
2275
2276<pre>
2277 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
2278</pre>
2279</div>
2280
2281
Robert Bocchino3a558662006-01-05 17:37:02 +0000[diff] [blame]2282<!-- _______________________________________________________________________ -->
2283<div class="doc_subsubsection">
2284 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2285</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2286
Robert Bocchino3a558662006-01-05 17:37:02 +0000[diff] [blame]2287<div class="doc_text">
2288
2289<h5>Syntax:</h5>
2290
2291<pre>
2292 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, uint &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2293</pre>
2294
2295<h5>Overview:</h5>
2296
2297<p>
2298The '<tt>extractelement</tt>' instruction extracts a single scalar
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]2299element from a packed vector at a specified index.
Robert Bocchino3a558662006-01-05 17:37:02 +0000[diff] [blame]2300</p>
2301
2302
2303<h5>Arguments:</h5>
2304
2305<p>
2306The first operand of an '<tt>extractelement</tt>' instruction is a
2307value of <a href="#t_packed">packed</a> type. The second operand is
2308an index indicating the position from which to extract the element.
2309The index may be a variable.</p>
2310
2311<h5>Semantics:</h5>
2312
2313<p>
2314The result is a scalar of the same type as the element type of
2315<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2316<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2317results are undefined.
2318</p>
2319
2320<h5>Example:</h5>
2321
2322<pre>
2323 %result = extractelement &lt;4 x int&gt; %vec, uint 0 <i>; yields int</i>
2324</pre>
2325</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2326
2327
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]2328<!-- _______________________________________________________________________ -->
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2329<div class="doc_subsubsection">
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]2330 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2331</div>
2332
2333<div class="doc_text">
2334
2335<h5>Syntax:</h5>
2336
2337<pre>
2338 &lt;result&gt; = insertelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt, uint &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2339</pre>
2340
2341<h5>Overview:</h5>
2342
2343<p>
2344The '<tt>insertelement</tt>' instruction inserts a scalar
2345element into a packed vector at a specified index.
2346</p>
2347
2348
2349<h5>Arguments:</h5>
2350
2351<p>
2352The first operand of an '<tt>insertelement</tt>' instruction is a
2353value of <a href="#t_packed">packed</a> type. The second operand is a
2354scalar value whose type must equal the element type of the first
2355operand. The third operand is an index indicating the position at
2356which to insert the value. The index may be a variable.</p>
2357
2358<h5>Semantics:</h5>
2359
2360<p>
2361The result is a packed vector of the same type as <tt>val</tt>. Its
2362element values are those of <tt>val</tt> except at position
2363<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2364exceeds the length of <tt>val</tt>, the results are undefined.
2365</p>
2366
2367<h5>Example:</h5>
2368
2369<pre>
2370 %result = insertelement &lt;4 x int&gt; %vec, int 1, uint 0 <i>; yields &lt;4 x int&gt;</i>
2371</pre>
2372</div>
2373
2374
2375<!-- _______________________________________________________________________ -->
2376<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2377 <a name="i_call">'<tt>call</tt>' Instruction</a>
2378</div>
2379
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2380<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2381
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2382<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2383<pre>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]2384 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] &lt;ty&gt;* &lt;fnptrval&gt;(&lt;param list&gt;)
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2385</pre>
2386
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2387<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2388
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2389<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2390
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2391<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2392
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2393<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2394
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2395<ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2396 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]2397 <p>The optional "tail" marker indicates whether the callee function accesses
2398 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2399 function call is eligible for tail call optimization. Note that calls may
2400 be marked "tail" even if they do not occur before a <a
2401 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2402 </li>
2403 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]2404 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
2405 convention</a> the call should use. If none is specified, the call defaults
2406 to using C calling conventions.
2407 </li>
2408 <li>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2409 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
2410 being invoked. The argument types must match the types implied by this
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2411 signature. This type can be omitted if the function is not varargs and
2412 if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2413 </li>
2414 <li>
2415 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
2416 be invoked. In most cases, this is a direct function invocation, but
2417 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2418 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2419 </li>
2420 <li>
2421 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +0000[diff] [blame]2422 function signature argument types. All arguments must be of
2423 <a href="#t_firstclass">first class</a> type. If the function signature
2424 indicates the function accepts a variable number of arguments, the extra
2425 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2426 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2427</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2428
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2429<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2430
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2431<p>The '<tt>call</tt>' instruction is used to cause control flow to
2432transfer to a specified function, with its incoming arguments bound to
2433the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
2434instruction in the called function, control flow continues with the
2435instruction after the function call, and the return value of the
2436function is bound to the result argument. This is a simpler case of
2437the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2438
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2439<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2440
2441<pre>
2442 %retval = call int %test(int %argc)
2443 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
2444 %X = tail call int %foo()
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]2445 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]2446</pre>
2447
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2448</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2449
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2450<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2451<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]2452 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2453</div>
2454
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2455<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2456
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2457<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2458
2459<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2460 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2461</pre>
2462
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2463<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2464
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2465<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2466the "variable argument" area of a function call. It is used to implement the
2467<tt>va_arg</tt> macro in C.</p>
2468
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2469<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2470
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2471<p>This instruction takes a <tt>va_list*</tt> value and the type of
2472the argument. It returns a value of the specified argument type and
Jeff Cohen25d4f7e2005-11-11 02:15:27 +0000[diff] [blame]2473increments the <tt>va_list</tt> to point to the next argument. Again, the
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2474actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2475
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2476<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2477
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2478<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
2479type from the specified <tt>va_list</tt> and causes the
2480<tt>va_list</tt> to point to the next argument. For more information,
2481see the variable argument handling <a href="#int_varargs">Intrinsic
2482Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2483
2484<p>It is legal for this instruction to be called in a function which does not
2485take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2486function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2487
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2488<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2489href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2490argument.</p>
2491
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]2492<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]2493
2494<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
2495
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2496</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2497
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2498<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2499<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
2500<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2501
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2502<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2503
2504<p>LLVM supports the notion of an "intrinsic function". These functions have
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2505well known names and semantics and are required to follow certain
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2506restrictions. Overall, these instructions represent an extension mechanism for
2507the LLVM language that does not require changing all of the transformations in
2508LLVM to add to the language (or the bytecode reader/writer, the parser,
2509etc...).</p>
2510
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2511<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
2512prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2513this. Intrinsic functions must always be external functions: you cannot define
2514the body of intrinsic functions. Intrinsic functions may only be used in call
2515or invoke instructions: it is illegal to take the address of an intrinsic
2516function. Additionally, because intrinsic functions are part of the LLVM
2517language, it is required that they all be documented here if any are added.</p>
2518
2519
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2520<p>To learn how to add an intrinsic function, please see the <a
Chris Lattner590cff32005-05-11 03:35:57 +0000[diff] [blame]2521href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2522</p>
2523
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2524</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2525
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2526<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2527<div class="doc_subsection">
2528 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
2529</div>
2530
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2531<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2532
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2533<p>Variable argument support is defined in LLVM with the <a
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]2534 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2535intrinsic functions. These functions are related to the similarly
2536named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2537
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2538<p>All of these functions operate on arguments that use a
2539target-specific value type "<tt>va_list</tt>". The LLVM assembly
2540language reference manual does not define what this type is, so all
2541transformations should be prepared to handle intrinsics with any type
2542used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2543
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2544<p>This example shows how the <a href="#i_vanext"><tt>vanext</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2545instruction and the variable argument handling intrinsic functions are
2546used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2547
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2548<pre>
2549int %test(int %X, ...) {
2550 ; Initialize variable argument processing
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2551 %ap = alloca sbyte*
2552 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2553
2554 ; Read a single integer argument
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2555 %tmp = va_arg sbyte** %ap, int
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2556
2557 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2558 %aq = alloca sbyte*
Andrew Lenharthd0a4c622005-06-22 20:38:11 +0000[diff] [blame]2559 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2560 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2561
2562 ; Stop processing of arguments.
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2563 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2564 ret int %tmp
2565}
2566</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2567</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2568
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2569<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2570<div class="doc_subsubsection">
2571 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
2572</div>
2573
2574
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2575<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2576<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2577<pre> declare void %llvm.va_start(&lt;va_list&gt;* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2578<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2579<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
2580<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
2581href="#i_va_arg">va_arg</a></tt>.</p>
2582
2583<h5>Arguments:</h5>
2584
2585<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
2586
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2587<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2588
2589<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
2590macro available in C. In a target-dependent way, it initializes the
2591<tt>va_list</tt> element the argument points to, so that the next call to
2592<tt>va_arg</tt> will produce the first variable argument passed to the function.
2593Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
2594last argument of the function, the compiler can figure that out.</p>
2595
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2596</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2597
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2598<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2599<div class="doc_subsubsection">
2600 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
2601</div>
2602
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2603<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2604<h5>Syntax:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2605<pre> declare void %llvm.va_end(&lt;va_list*&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2606<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2607<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
2608which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
2609or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2610<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2611<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2612<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2613<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]2614macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
2615Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
2616 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
2617with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2618</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2619
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2620<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2621<div class="doc_subsubsection">
2622 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
2623</div>
2624
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2625<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2626
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2627<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2628
2629<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2630 declare void %llvm.va_copy(&lt;va_list&gt;* &lt;destarglist&gt;,
Andrew Lenharthd0a4c622005-06-22 20:38:11 +0000[diff] [blame]2631 &lt;va_list&gt;* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2632</pre>
2633
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2634<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2635
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2636<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
2637the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2638
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2639<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2640
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2641<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +0000[diff] [blame]2642The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2643
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2644
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]2645<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2646
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]2647<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
2648available in C. In a target-dependent way, it copies the source
2649<tt>va_list</tt> element into the destination list. This intrinsic is necessary
2650because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2651arbitrarily complex and require memory allocation, for example.</p>
2652
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2653</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]2654
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]2655<!-- ======================================================================= -->
2656<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2657 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
2658</div>
2659
2660<div class="doc_text">
2661
2662<p>
2663LLVM support for <a href="GarbageCollection.html">Accurate Garbage
2664Collection</a> requires the implementation and generation of these intrinsics.
2665These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
2666stack</a>, as well as garbage collector implementations that require <a
2667href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
2668Front-ends for type-safe garbage collected languages should generate these
2669intrinsics to make use of the LLVM garbage collectors. For more details, see <a
2670href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
2671</p>
2672</div>
2673
2674<!-- _______________________________________________________________________ -->
2675<div class="doc_subsubsection">
2676 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
2677</div>
2678
2679<div class="doc_text">
2680
2681<h5>Syntax:</h5>
2682
2683<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]2684 declare void %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2685</pre>
2686
2687<h5>Overview:</h5>
2688
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2689<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]2690the code generator, and allows some metadata to be associated with it.</p>
2691
2692<h5>Arguments:</h5>
2693
2694<p>The first argument specifies the address of a stack object that contains the
2695root pointer. The second pointer (which must be either a constant or a global
2696value address) contains the meta-data to be associated with the root.</p>
2697
2698<h5>Semantics:</h5>
2699
2700<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
2701location. At compile-time, the code generator generates information to allow
2702the runtime to find the pointer at GC safe points.
2703</p>
2704
2705</div>
2706
2707
2708<!-- _______________________________________________________________________ -->
2709<div class="doc_subsubsection">
2710 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
2711</div>
2712
2713<div class="doc_text">
2714
2715<h5>Syntax:</h5>
2716
2717<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]2718 declare sbyte* %llvm.gcread(sbyte** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2719</pre>
2720
2721<h5>Overview:</h5>
2722
2723<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
2724locations, allowing garbage collector implementations that require read
2725barriers.</p>
2726
2727<h5>Arguments:</h5>
2728
2729<p>The argument is the address to read from, which should be an address
2730allocated from the garbage collector.</p>
2731
2732<h5>Semantics:</h5>
2733
2734<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
2735instruction, but may be replaced with substantially more complex code by the
2736garbage collector runtime, as needed.</p>
2737
2738</div>
2739
2740
2741<!-- _______________________________________________________________________ -->
2742<div class="doc_subsubsection">
2743 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
2744</div>
2745
2746<div class="doc_text">
2747
2748<h5>Syntax:</h5>
2749
2750<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]2751 declare void %llvm.gcwrite(sbyte* %P1, sbyte** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]2752</pre>
2753
2754<h5>Overview:</h5>
2755
2756<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
2757locations, allowing garbage collector implementations that require write
2758barriers (such as generational or reference counting collectors).</p>
2759
2760<h5>Arguments:</h5>
2761
2762<p>The first argument is the reference to store, and the second is the heap
2763location to store to.</p>
2764
2765<h5>Semantics:</h5>
2766
2767<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
2768instruction, but may be replaced with substantially more complex code by the
2769garbage collector runtime, as needed.</p>
2770
2771</div>
2772
2773
2774
2775<!-- ======================================================================= -->
2776<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2777 <a name="int_codegen">Code Generator Intrinsics</a>
2778</div>
2779
2780<div class="doc_text">
2781<p>
2782These intrinsics are provided by LLVM to expose special features that may only
2783be implemented with code generator support.
2784</p>
2785
2786</div>
2787
2788<!-- _______________________________________________________________________ -->
2789<div class="doc_subsubsection">
2790 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
2791</div>
2792
2793<div class="doc_text">
2794
2795<h5>Syntax:</h5>
2796<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +0000[diff] [blame]2797 declare sbyte *%llvm.returnaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2798</pre>
2799
2800<h5>Overview:</h5>
2801
2802<p>
2803The '<tt>llvm.returnaddress</tt>' intrinsic returns a target-specific value
2804indicating the return address of the current function or one of its callers.
2805</p>
2806
2807<h5>Arguments:</h5>
2808
2809<p>
2810The argument to this intrinsic indicates which function to return the address
2811for. Zero indicates the calling function, one indicates its caller, etc. The
2812argument is <b>required</b> to be a constant integer value.
2813</p>
2814
2815<h5>Semantics:</h5>
2816
2817<p>
2818The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
2819the return address of the specified call frame, or zero if it cannot be
2820identified. The value returned by this intrinsic is likely to be incorrect or 0
2821for arguments other than zero, so it should only be used for debugging purposes.
2822</p>
2823
2824<p>
2825Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +0000[diff] [blame]2826aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2827source-language caller.
2828</p>
2829</div>
2830
2831
2832<!-- _______________________________________________________________________ -->
2833<div class="doc_subsubsection">
2834 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
2835</div>
2836
2837<div class="doc_text">
2838
2839<h5>Syntax:</h5>
2840<pre>
Chris Lattnerfcf39d42006-01-13 01:20:27 +0000[diff] [blame]2841 declare sbyte *%llvm.frameaddress(uint &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2842</pre>
2843
2844<h5>Overview:</h5>
2845
2846<p>
2847The '<tt>llvm.frameaddress</tt>' intrinsic returns the target-specific frame
2848pointer value for the specified stack frame.
2849</p>
2850
2851<h5>Arguments:</h5>
2852
2853<p>
2854The argument to this intrinsic indicates which function to return the frame
2855pointer for. Zero indicates the calling function, one indicates its caller,
2856etc. The argument is <b>required</b> to be a constant integer value.
2857</p>
2858
2859<h5>Semantics:</h5>
2860
2861<p>
2862The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
2863the frame address of the specified call frame, or zero if it cannot be
2864identified. The value returned by this intrinsic is likely to be incorrect or 0
2865for arguments other than zero, so it should only be used for debugging purposes.
2866</p>
2867
2868<p>
2869Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +0000[diff] [blame]2870aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]2871source-language caller.
2872</p>
2873</div>
2874
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]2875<!-- _______________________________________________________________________ -->
2876<div class="doc_subsubsection">
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]2877 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
2878</div>
2879
2880<div class="doc_text">
2881
2882<h5>Syntax:</h5>
2883<pre>
2884 declare sbyte *%llvm.stacksave()
2885</pre>
2886
2887<h5>Overview:</h5>
2888
2889<p>
2890The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
2891the function stack, for use with <a href="#i_stackrestore">
2892<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
2893features like scoped automatic variable sized arrays in C99.
2894</p>
2895
2896<h5>Semantics:</h5>
2897
2898<p>
2899This intrinsic returns a opaque pointer value that can be passed to <a
2900href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
2901<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
2902<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
2903state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
2904practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
2905that were allocated after the <tt>llvm.stacksave</tt> was executed.
2906</p>
2907
2908</div>
2909
2910<!-- _______________________________________________________________________ -->
2911<div class="doc_subsubsection">
2912 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
2913</div>
2914
2915<div class="doc_text">
2916
2917<h5>Syntax:</h5>
2918<pre>
2919 declare void %llvm.stackrestore(sbyte* %ptr)
2920</pre>
2921
2922<h5>Overview:</h5>
2923
2924<p>
2925The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
2926the function stack to the state it was in when the corresponding <a
2927href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
2928useful for implementing language features like scoped automatic variable sized
2929arrays in C99.
2930</p>
2931
2932<h5>Semantics:</h5>
2933
2934<p>
2935See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
2936</p>
2937
2938</div>
2939
2940
2941<!-- _______________________________________________________________________ -->
2942<div class="doc_subsubsection">
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]2943 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
2944</div>
2945
2946<div class="doc_text">
2947
2948<h5>Syntax:</h5>
2949<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]2950 declare void %llvm.prefetch(sbyte * &lt;address&gt;,
2951 uint &lt;rw&gt;, uint &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]2952</pre>
2953
2954<h5>Overview:</h5>
2955
2956
2957<p>
2958The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2959a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
2960no
2961effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +0000[diff] [blame]2962characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]2963</p>
2964
2965<h5>Arguments:</h5>
2966
2967<p>
2968<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
2969determining if the fetch should be for a read (0) or write (1), and
2970<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +0000[diff] [blame]2971locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]2972<tt>locality</tt> arguments must be constant integers.
2973</p>
2974
2975<h5>Semantics:</h5>
2976
2977<p>
2978This intrinsic does not modify the behavior of the program. In particular,
2979prefetches cannot trap and do not produce a value. On targets that support this
2980intrinsic, the prefetch can provide hints to the processor cache for better
2981performance.
2982</p>
2983
2984</div>
2985
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]2986<!-- _______________________________________________________________________ -->
2987<div class="doc_subsubsection">
2988 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
2989</div>
2990
2991<div class="doc_text">
2992
2993<h5>Syntax:</h5>
2994<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]2995 declare void %llvm.pcmarker( uint &lt;id&gt; )
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]2996</pre>
2997
2998<h5>Overview:</h5>
2999
3000
3001<p>
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3002The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
3003(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]3004code to simulators and other tools. The method is target specific, but it is
3005expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +0000[diff] [blame]3006The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +0000[diff] [blame]3007after optimizations. It is possible that the presence of a marker will inhibit
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]3008optimizations. The intended use is to be inserted after optmizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3009correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]3010</p>
3011
3012<h5>Arguments:</h5>
3013
3014<p>
3015<tt>id</tt> is a numerical id identifying the marker.
3016</p>
3017
3018<h5>Semantics:</h5>
3019
3020<p>
3021This intrinsic does not modify the behavior of the program. Backends that do not
3022support this intrinisic may ignore it.
3023</p>
3024
3025</div>
3026
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]3027<!-- _______________________________________________________________________ -->
3028<div class="doc_subsubsection">
3029 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
3030</div>
3031
3032<div class="doc_text">
3033
3034<h5>Syntax:</h5>
3035<pre>
3036 declare ulong %llvm.readcyclecounter( )
3037</pre>
3038
3039<h5>Overview:</h5>
3040
3041
3042<p>
3043The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
3044counter register (or similar low latency, high accuracy clocks) on those targets
3045that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
3046As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
3047should only be used for small timings.
3048</p>
3049
3050<h5>Semantics:</h5>
3051
3052<p>
3053When directly supported, reading the cycle counter should not modify any memory.
3054Implementations are allowed to either return a application specific value or a
3055system wide value. On backends without support, this is lowered to a constant 0.
3056</p>
3057
3058</div>
3059
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]3060
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]3061<!-- ======================================================================= -->
3062<div class="doc_subsection">
3063 <a name="int_os">Operating System Intrinsics</a>
3064</div>
3065
3066<div class="doc_text">
3067<p>
3068These intrinsics are provided by LLVM to support the implementation of
3069operating system level code.
3070</p>
3071
3072</div>
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3073
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3074<!-- _______________________________________________________________________ -->
3075<div class="doc_subsubsection">
3076 <a name="i_readport">'<tt>llvm.readport</tt>' Intrinsic</a>
3077</div>
3078
3079<div class="doc_text">
3080
3081<h5>Syntax:</h5>
3082<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]3083 declare &lt;integer type&gt; %llvm.readport (&lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3084</pre>
3085
3086<h5>Overview:</h5>
3087
3088<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]3089The '<tt>llvm.readport</tt>' intrinsic reads data from the specified hardware
3090I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3091</p>
3092
3093<h5>Arguments:</h5>
3094
3095<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]3096The argument to this intrinsic indicates the hardware I/O address from which
3097to read the data. The address is in the hardware I/O address namespace (as
3098opposed to being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3099</p>
3100
3101<h5>Semantics:</h5>
3102
3103<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]3104The '<tt>llvm.readport</tt>' intrinsic reads data from the hardware I/O port
3105specified by <i>address</i> and returns the value. The address and return
3106value must be integers, but the size is dependent upon the platform upon which
3107the program is code generated. For example, on x86, the address must be an
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000[diff] [blame]3108unsigned 16-bit value, and the return value must be 8, 16, or 32 bits.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3109</p>
3110
3111</div>
3112
3113<!-- _______________________________________________________________________ -->
3114<div class="doc_subsubsection">
3115 <a name="i_writeport">'<tt>llvm.writeport</tt>' Intrinsic</a>
3116</div>
3117
3118<div class="doc_text">
3119
3120<h5>Syntax:</h5>
3121<pre>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3122 call void (&lt;integer type&gt;, &lt;integer type&gt;)*
3123 %llvm.writeport (&lt;integer type&gt; &lt;value&gt;,
3124 &lt;integer type&gt; &lt;address&gt;)
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3125</pre>
3126
3127<h5>Overview:</h5>
3128
3129<p>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]3130The '<tt>llvm.writeport</tt>' intrinsic writes data to the specified hardware
3131I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3132</p>
3133
3134<h5>Arguments:</h5>
3135
3136<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3137The first argument is the value to write to the I/O port.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3138</p>
3139
3140<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3141The second argument indicates the hardware I/O address to which data should be
3142written. The address is in the hardware I/O address namespace (as opposed to
3143being a memory location for memory mapped I/O).
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3144</p>
3145
3146<h5>Semantics:</h5>
3147
3148<p>
3149The '<tt>llvm.writeport</tt>' intrinsic writes <i>value</i> to the I/O port
3150specified by <i>address</i>. The address and value must be integers, but the
3151size is dependent upon the platform upon which the program is code generated.
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000[diff] [blame]3152For example, on x86, the address must be an unsigned 16-bit value, and the
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]3153value written must be 8, 16, or 32 bits in length.
John Criswellcfd3bac2004-04-09 15:23:37 +0000[diff] [blame]3154</p>
3155
3156</div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]3157
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3158<!-- _______________________________________________________________________ -->
3159<div class="doc_subsubsection">
3160 <a name="i_readio">'<tt>llvm.readio</tt>' Intrinsic</a>
3161</div>
3162
3163<div class="doc_text">
3164
3165<h5>Syntax:</h5>
3166<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]3167 declare &lt;result&gt; %llvm.readio (&lt;ty&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3168</pre>
3169
3170<h5>Overview:</h5>
3171
3172<p>
3173The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
3174address.
3175</p>
3176
3177<h5>Arguments:</h5>
3178
3179<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3180The argument to this intrinsic is a pointer indicating the memory address from
3181which to read the data. The data must be a
3182<a href="#t_firstclass">first class</a> type.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3183</p>
3184
3185<h5>Semantics:</h5>
3186
3187<p>
3188The '<tt>llvm.readio</tt>' intrinsic reads data from a memory mapped I/O
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3189location specified by <i>pointer</i> and returns the value. The argument must
3190be a pointer, and the return value must be a
3191<a href="#t_firstclass">first class</a> type. However, certain architectures
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000[diff] [blame]3192may not support I/O on all first class types. For example, 32-bit processors
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3193may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3194</p>
3195
3196<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3197This intrinsic enforces an in-order memory model for llvm.readio and
3198llvm.writeio calls on machines that use dynamic scheduling. Dynamically
3199scheduled processors may execute loads and stores out of order, re-ordering at
3200run time accesses to memory mapped I/O registers. Using these intrinsics
3201ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3202</p>
3203
3204</div>
3205
3206<!-- _______________________________________________________________________ -->
3207<div class="doc_subsubsection">
3208 <a name="i_writeio">'<tt>llvm.writeio</tt>' Intrinsic</a>
3209</div>
3210
3211<div class="doc_text">
3212
3213<h5>Syntax:</h5>
3214<pre>
Reid Spencera8d451e2005-04-26 20:50:44 +0000[diff] [blame]3215 declare void %llvm.writeio (&lt;ty1&gt; &lt;value&gt;, &lt;ty2&gt; * &lt;pointer&gt;)
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3216</pre>
3217
3218<h5>Overview:</h5>
3219
3220<p>
3221The '<tt>llvm.writeio</tt>' intrinsic writes data to the specified memory
3222mapped I/O address.
3223</p>
3224
3225<h5>Arguments:</h5>
3226
3227<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3228The first argument is the value to write to the memory mapped I/O location.
3229The second argument is a pointer indicating the memory address to which the
3230data should be written.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3231</p>
3232
3233<h5>Semantics:</h5>
3234
3235<p>
3236The '<tt>llvm.writeio</tt>' intrinsic writes <i>value</i> to the memory mapped
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3237I/O address specified by <i>pointer</i>. The value must be a
3238<a href="#t_firstclass">first class</a> type. However, certain architectures
Misha Brukmancfa87bc2005-04-22 18:02:52 +0000[diff] [blame]3239may not support I/O on all first class types. For example, 32-bit processors
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3240may only support I/O on data types that are 32 bits or less.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3241</p>
3242
3243<p>
John Criswell96db6fc2004-04-12 16:33:19 +0000[diff] [blame]3244This intrinsic enforces an in-order memory model for llvm.readio and
3245llvm.writeio calls on machines that use dynamic scheduling. Dynamically
3246scheduled processors may execute loads and stores out of order, re-ordering at
3247run time accesses to memory mapped I/O registers. Using these intrinsics
3248ensures that accesses to memory mapped I/O registers occur in program order.
John Criswell183402a2004-04-12 15:02:16 +0000[diff] [blame]3249</p>
3250
3251</div>
3252
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]3253<!-- ======================================================================= -->
3254<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3255 <a name="int_libc">Standard C Library Intrinsics</a>
3256</div>
3257
3258<div class="doc_text">
3259<p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]3260LLVM provides intrinsics for a few important standard C library functions.
3261These intrinsics allow source-language front-ends to pass information about the
3262alignment of the pointer arguments to the code generator, providing opportunity
3263for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3264</p>
3265
3266</div>
3267
3268<!-- _______________________________________________________________________ -->
3269<div class="doc_subsubsection">
3270 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
3271</div>
3272
3273<div class="doc_text">
3274
3275<h5>Syntax:</h5>
3276<pre>
Reid Spencerd4622352005-04-26 20:41:16 +0000[diff] [blame]3277 declare void %llvm.memcpy(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3278 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3279</pre>
3280
3281<h5>Overview:</h5>
3282
3283<p>
3284The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
3285location to the destination location.
3286</p>
3287
3288<p>
3289Note that, unlike the standard libc function, the <tt>llvm.memcpy</tt> intrinsic
3290does not return a value, and takes an extra alignment argument.
3291</p>
3292
3293<h5>Arguments:</h5>
3294
3295<p>
3296The first argument is a pointer to the destination, the second is a pointer to
3297the source. The third argument is an (arbitrarily sized) integer argument
3298specifying the number of bytes to copy, and the fourth argument is the alignment
3299of the source and destination locations.
3300</p>
3301
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]3302<p>
3303If the call to this intrinisic has an alignment value that is not 0 or 1, then
3304the caller guarantees that the size of the copy is a multiple of the alignment
3305and that both the source and destination pointers are aligned to that boundary.
3306</p>
3307
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3308<h5>Semantics:</h5>
3309
3310<p>
3311The '<tt>llvm.memcpy</tt>' intrinsic copies a block of memory from the source
3312location to the destination location, which are not allowed to overlap. It
3313copies "len" bytes of memory over. If the argument is known to be aligned to
3314some boundary, this can be specified as the fourth argument, otherwise it should
3315be set to 0 or 1.
3316</p>
3317</div>
3318
3319
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]3320<!-- _______________________________________________________________________ -->
3321<div class="doc_subsubsection">
3322 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
3323</div>
3324
3325<div class="doc_text">
3326
3327<h5>Syntax:</h5>
3328<pre>
Reid Spencerd4622352005-04-26 20:41:16 +0000[diff] [blame]3329 declare void %llvm.memmove(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
3330 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]3331</pre>
3332
3333<h5>Overview:</h5>
3334
3335<p>
3336The '<tt>llvm.memmove</tt>' intrinsic moves a block of memory from the source
3337location to the destination location. It is similar to the '<tt>llvm.memcpy</tt>'
3338intrinsic but allows the two memory locations to overlap.
3339</p>
3340
3341<p>
3342Note that, unlike the standard libc function, the <tt>llvm.memmove</tt> intrinsic
3343does not return a value, and takes an extra alignment argument.
3344</p>
3345
3346<h5>Arguments:</h5>
3347
3348<p>
3349The first argument is a pointer to the destination, the second is a pointer to
3350the source. The third argument is an (arbitrarily sized) integer argument
3351specifying the number of bytes to copy, and the fourth argument is the alignment
3352of the source and destination locations.
3353</p>
3354
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]3355<p>
3356If the call to this intrinisic has an alignment value that is not 0 or 1, then
3357the caller guarantees that the size of the copy is a multiple of the alignment
3358and that both the source and destination pointers are aligned to that boundary.
3359</p>
3360
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]3361<h5>Semantics:</h5>
3362
3363<p>
3364The '<tt>llvm.memmove</tt>' intrinsic copies a block of memory from the source
3365location to the destination location, which may overlap. It
3366copies "len" bytes of memory over. If the argument is known to be aligned to
3367some boundary, this can be specified as the fourth argument, otherwise it should
3368be set to 0 or 1.
3369</p>
3370</div>
3371
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3372
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]3373<!-- _______________________________________________________________________ -->
3374<div class="doc_subsubsection">
3375 <a name="i_memset">'<tt>llvm.memset</tt>' Intrinsic</a>
3376</div>
3377
3378<div class="doc_text">
3379
3380<h5>Syntax:</h5>
3381<pre>
Reid Spencerd4622352005-04-26 20:41:16 +0000[diff] [blame]3382 declare void %llvm.memset(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
3383 uint &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]3384</pre>
3385
3386<h5>Overview:</h5>
3387
3388<p>
3389The '<tt>llvm.memset</tt>' intrinsic fills a block of memory with a particular
3390byte value.
3391</p>
3392
3393<p>
3394Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
3395does not return a value, and takes an extra alignment argument.
3396</p>
3397
3398<h5>Arguments:</h5>
3399
3400<p>
3401The first argument is a pointer to the destination to fill, the second is the
3402byte value to fill it with, the third argument is an (arbitrarily sized) integer
3403argument specifying the number of bytes to fill, and the fourth argument is the
3404known alignment of destination location.
3405</p>
3406
3407<p>
3408If the call to this intrinisic has an alignment value that is not 0 or 1, then
3409the caller guarantees that the size of the copy is a multiple of the alignment
3410and that the destination pointer is aligned to that boundary.
3411</p>
3412
3413<h5>Semantics:</h5>
3414
3415<p>
3416The '<tt>llvm.memset</tt>' intrinsic fills "len" bytes of memory starting at the
3417destination location. If the argument is known to be aligned to some boundary,
3418this can be specified as the fourth argument, otherwise it should be set to 0 or
34191.
3420</p>
3421</div>
3422
3423
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]3424<!-- _______________________________________________________________________ -->
3425<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3426 <a name="i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a>
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +0000[diff] [blame]3427</div>
3428
3429<div class="doc_text">
3430
3431<h5>Syntax:</h5>
3432<pre>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3433 declare bool %llvm.isunordered.f32(float Val1, float Val2)
3434 declare bool %llvm.isunordered.f64(double Val1, double Val2)
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +0000[diff] [blame]3435</pre>
3436
3437<h5>Overview:</h5>
3438
3439<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3440The '<tt>llvm.isunordered</tt>' intrinsics return true if either or both of the
Alkis Evlogimenos26bbe932004-06-13 01:16:15 +0000[diff] [blame]3441specified floating point values is a NAN.
3442</p>
3443
3444<h5>Arguments:</h5>
3445
3446<p>
3447The arguments are floating point numbers of the same type.
3448</p>
3449
3450<h5>Semantics:</h5>
3451
3452<p>
3453If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
3454false.
3455</p>
3456</div>
3457
3458
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]3459<!-- _______________________________________________________________________ -->
3460<div class="doc_subsubsection">
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3461 <a name="i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]3462</div>
3463
3464<div class="doc_text">
3465
3466<h5>Syntax:</h5>
3467<pre>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3468 declare double %llvm.sqrt.f32(float Val)
3469 declare double %llvm.sqrt.f64(double Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]3470</pre>
3471
3472<h5>Overview:</h5>
3473
3474<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3475The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]3476returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
3477<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
3478negative numbers (which allows for better optimization).
3479</p>
3480
3481<h5>Arguments:</h5>
3482
3483<p>
3484The argument and return value are floating point numbers of the same type.
3485</p>
3486
3487<h5>Semantics:</h5>
3488
3489<p>
3490This function returns the sqrt of the specified operand if it is a positive
3491floating point number.
3492</p>
3493</div>
3494
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3495<!-- ======================================================================= -->
3496<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]3497 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3498</div>
3499
3500<div class="doc_text">
3501<p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]3502LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3503These allow efficient code generation for some algorithms.
3504</p>
3505
3506</div>
3507
3508<!-- _______________________________________________________________________ -->
3509<div class="doc_subsubsection">
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]3510 <a name="i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
3511</div>
3512
3513<div class="doc_text">
3514
3515<h5>Syntax:</h5>
3516<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3517 declare ushort %llvm.bswap.i16(ushort &lt;id&gt;)
3518 declare uint %llvm.bswap.i32(uint &lt;id&gt;)
3519 declare ulong %llvm.bswap.i64(ulong &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]3520</pre>
3521
3522<h5>Overview:</h5>
3523
3524<p>
3525The '<tt>llvm.bwsap</tt>' family of intrinsics is used to byteswap a 16, 32 or
352664 bit quantity. These are useful for performing operations on data that is not
3527in the target's native byte order.
3528</p>
3529
3530<h5>Semantics:</h5>
3531
3532<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3533The <tt>llvm.bswap.16</tt> intrinsic returns a ushort value that has the high and low
3534byte of the input ushort swapped. Similarly, the <tt>llvm.bswap.i32</tt> intrinsic
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]3535returns a uint value that has the four bytes of the input uint swapped, so that
3536if the input bytes are numbered 0, 1, 2, 3 then the returned uint will have its
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3537bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i64</tt> intrinsic extends this concept
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]3538to 64 bits.
3539</p>
3540
3541</div>
3542
3543<!-- _______________________________________________________________________ -->
3544<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3545 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3546</div>
3547
3548<div class="doc_text">
3549
3550<h5>Syntax:</h5>
3551<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3552 declare ubyte %llvm.ctpop.i8 (ubyte &lt;src&gt;)
3553 declare ushort %llvm.ctpop.i16(ushort &lt;src&gt;)
3554 declare uint %llvm.ctpop.i32(uint &lt;src&gt;)
3555 declare ulong %llvm.ctpop.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3556</pre>
3557
3558<h5>Overview:</h5>
3559
3560<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3561The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
3562value.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3563</p>
3564
3565<h5>Arguments:</h5>
3566
3567<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]3568The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3569unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3570</p>
3571
3572<h5>Semantics:</h5>
3573
3574<p>
3575The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
3576</p>
3577</div>
3578
3579<!-- _______________________________________________________________________ -->
3580<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]3581 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3582</div>
3583
3584<div class="doc_text">
3585
3586<h5>Syntax:</h5>
3587<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3588 declare ubyte %llvm.ctlz.i8 (ubyte &lt;src&gt;)
3589 declare ushort %llvm.ctlz.i16(ushort &lt;src&gt;)
3590 declare uint %llvm.ctlz.i32(uint &lt;src&gt;)
3591 declare ulong %llvm.ctlz.i64(ulong &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3592</pre>
3593
3594<h5>Overview:</h5>
3595
3596<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3597The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
3598leading zeros in a variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3599</p>
3600
3601<h5>Arguments:</h5>
3602
3603<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]3604The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3605unsigned integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3606</p>
3607
3608<h5>Semantics:</h5>
3609
3610<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]3611The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
3612in a variable. If the src == 0 then the result is the size in bits of the type
3613of src. For example, <tt>llvm.cttz(int 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]3614</p>
3615</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]3616
3617
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]3618
3619<!-- _______________________________________________________________________ -->
3620<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]3621 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]3622</div>
3623
3624<div class="doc_text">
3625
3626<h5>Syntax:</h5>
3627<pre>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3628 declare ubyte %llvm.cttz.i8 (ubyte &lt;src&gt;)
3629 declare ushort %llvm.cttz.i16(ushort &lt;src&gt;)
3630 declare uint %llvm.cttz.i32(uint &lt;src&gt;)
3631 declare ulong %llvm.cttz.i64(ulong &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]3632</pre>
3633
3634<h5>Overview:</h5>
3635
3636<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]3637The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
3638trailing zeros.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]3639</p>
3640
3641<h5>Arguments:</h5>
3642
3643<p>
3644The only argument is the value to be counted. The argument may be of any
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]3645unsigned integer type. The return type must match the argument type.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]3646</p>
3647
3648<h5>Semantics:</h5>
3649
3650<p>
3651The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
3652in a variable. If the src == 0 then the result is the size in bits of the type
3653of src. For example, <tt>llvm.cttz(2) = 1</tt>.
3654</p>
3655</div>
3656
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3657<!-- ======================================================================= -->
3658<div class="doc_subsection">
3659 <a name="int_debugger">Debugger Intrinsics</a>
3660</div>
3661
3662<div class="doc_text">
3663<p>
3664The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
3665are described in the <a
3666href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
3667Debugging</a> document.
3668</p>
3669</div>
3670
3671
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3672<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3673<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]3674<address>
3675 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
3676 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
3677 <a href="http://validator.w3.org/check/referer"><img
3678 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
3679
3680 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
3681 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a><br>
3682 Last modified: $Date$
3683</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3684</body>
3685</html>