blob: 1c744d2a6fee690b1080bf719917ce66df80b27f [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman9d0919f2003-11-08 01:05:38 +00003<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +00006 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
8 <meta name="description"
9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman9d0919f2003-11-08 01:05:38 +000010 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +000012
Misha Brukman9d0919f2003-11-08 01:05:38 +000013<body>
Chris Lattnerd7923912004-05-23 21:06:01 +000014
Chris Lattner261efe92003-11-25 01:02:51 +000015<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner00950542001-06-06 20:29:01 +000016<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +000017 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000020 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Chris Lattnere5d947b2004-12-09 16:36:40 +000023 <li><a href="#linkage">Linkage Types</a></li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +000024 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000025 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +000026 <li><a href="#functionstructure">Functions</a></li>
Anton Korobeynikovc6c98af2007-04-29 18:02:48 +000027 <li><a href="#aliasstructure">Aliases</a>
Reid Spencerca86e162006-12-31 07:07:53 +000028 <li><a href="#paramattrs">Parameter Attributes</a></li>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +000029 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +000030 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencerde151942007-02-19 23:54:10 +000031 <li><a href="#datalayout">Data Layout</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000032 </ol>
33 </li>
Chris Lattner00950542001-06-06 20:29:01 +000034 <li><a href="#typesystem">Type System</a>
35 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +000036 <li><a href="#t_classifications">Type Classifications</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +000037 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +000038 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +000039 <li><a href="#t_floating">Floating Point Types</a></li>
40 <li><a href="#t_void">Void Type</a></li>
41 <li><a href="#t_label">Label Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000042 </ol>
43 </li>
Chris Lattner00950542001-06-06 20:29:01 +000044 <li><a href="#t_derived">Derived Types</a>
45 <ol>
Chris Lattnerb9488a62007-12-18 06:18:21 +000046 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000047 <li><a href="#t_array">Array Type</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000048 <li><a href="#t_function">Function Type</a></li>
49 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000050 <li><a href="#t_struct">Structure Type</a></li>
Andrew Lenharth75e10682006-12-08 17:13:00 +000051 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Reid Spencer485bad12007-02-15 03:07:05 +000052 <li><a href="#t_vector">Vector Type</a></li>
Chris Lattner69c11bb2005-04-25 17:34:15 +000053 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000054 </ol>
55 </li>
56 </ol>
57 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +000058 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +000059 <ol>
60 <li><a href="#simpleconstants">Simple Constants</a>
61 <li><a href="#aggregateconstants">Aggregate Constants</a>
62 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
63 <li><a href="#undefvalues">Undefined Values</a>
64 <li><a href="#constantexprs">Constant Expressions</a>
65 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +000066 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +000067 <li><a href="#othervalues">Other Values</a>
68 <ol>
69 <li><a href="#inlineasm">Inline Assembler Expressions</a>
70 </ol>
71 </li>
Chris Lattner00950542001-06-06 20:29:01 +000072 <li><a href="#instref">Instruction Reference</a>
73 <ol>
74 <li><a href="#terminators">Terminator Instructions</a>
75 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000076 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
77 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000078 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
79 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000080 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +000081 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000082 </ol>
83 </li>
Chris Lattner00950542001-06-06 20:29:01 +000084 <li><a href="#binaryops">Binary Operations</a>
85 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +000086 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
87 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
88 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +000089 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
90 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
91 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +000092 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
93 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
94 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000095 </ol>
96 </li>
Chris Lattner00950542001-06-06 20:29:01 +000097 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
98 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +000099 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
100 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
101 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000102 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000103 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000104 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000105 </ol>
106 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000107 <li><a href="#vectorops">Vector Operations</a>
108 <ol>
109 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
110 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
111 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000112 </ol>
113 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000114 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000115 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000116 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
117 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
118 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000119 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
120 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
121 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000122 </ol>
123 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000124 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000125 <ol>
126 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
127 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
128 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
129 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
130 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000131 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
132 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
133 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
134 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000135 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
136 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000137 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000138 </ol>
Chris Lattner00950542001-06-06 20:29:01 +0000139 <li><a href="#otherops">Other Operations</a>
140 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000141 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
142 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000143 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000144 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000145 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000146 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000147 <li><a href="#i_getresult">'<tt>getresult</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000148 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000149 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000150 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000151 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000152 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000153 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000154 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
155 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000156 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
157 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
158 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000159 </ol>
160 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000161 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
162 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000163 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
164 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
165 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000166 </ol>
167 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000168 <li><a href="#int_codegen">Code Generator Intrinsics</a>
169 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000170 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
171 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
172 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
173 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
174 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
175 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
176 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000177 </ol>
178 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000179 <li><a href="#int_libc">Standard C Library Intrinsics</a>
180 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000181 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
182 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
183 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
184 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
185 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000186 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
187 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
188 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000189 </ol>
190 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000191 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000192 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000193 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000194 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
195 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
196 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Reid Spencerf86037f2007-04-11 23:23:49 +0000197 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
198 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000199 </ol>
200 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000201 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000202 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000203 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000204 <ol>
205 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000206 </ol>
207 </li>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000208 <li><a href="#int_atomics">Atomic intrinsics</a>
209 <ol>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000210 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
211 <li><a href="#int_atomic_lcs"><tt>llvm.atomic.lcs</tt></a></li>
212 <li><a href="#int_atomic_las"><tt>llvm.atomic.las</tt></a></li>
213 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000214 </ol>
215 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000216 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000217 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000218 <li><a href="#int_var_annotation">
Tanya Lattner91d0b882007-09-22 00:01:26 +0000219 <tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000220 <li><a href="#int_annotation">
Tanya Lattner91d0b882007-09-22 00:01:26 +0000221 <tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000222 <li><a href="#int_trap">
223 <tt>llvm.trap</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000224 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000225 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000226 </ol>
227 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000228</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000229
230<div class="doc_author">
231 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
232 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000233</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000234
Chris Lattner00950542001-06-06 20:29:01 +0000235<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000236<div class="doc_section"> <a name="abstract">Abstract </a></div>
237<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000238
Misha Brukman9d0919f2003-11-08 01:05:38 +0000239<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000240<p>This document is a reference manual for the LLVM assembly language.
241LLVM is an SSA based representation that provides type safety,
242low-level operations, flexibility, and the capability of representing
243'all' high-level languages cleanly. It is the common code
244representation used throughout all phases of the LLVM compilation
245strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000246</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000247
Chris Lattner00950542001-06-06 20:29:01 +0000248<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000249<div class="doc_section"> <a name="introduction">Introduction</a> </div>
250<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000251
Misha Brukman9d0919f2003-11-08 01:05:38 +0000252<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000253
Chris Lattner261efe92003-11-25 01:02:51 +0000254<p>The LLVM code representation is designed to be used in three
Gabor Greif04367bf2007-07-06 22:07:22 +0000255different forms: as an in-memory compiler IR, as an on-disk bitcode
Chris Lattner261efe92003-11-25 01:02:51 +0000256representation (suitable for fast loading by a Just-In-Time compiler),
257and as a human readable assembly language representation. This allows
258LLVM to provide a powerful intermediate representation for efficient
259compiler transformations and analysis, while providing a natural means
260to debug and visualize the transformations. The three different forms
261of LLVM are all equivalent. This document describes the human readable
262representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000263
John Criswellc1f786c2005-05-13 22:25:59 +0000264<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000265while being expressive, typed, and extensible at the same time. It
266aims to be a "universal IR" of sorts, by being at a low enough level
267that high-level ideas may be cleanly mapped to it (similar to how
268microprocessors are "universal IR's", allowing many source languages to
269be mapped to them). By providing type information, LLVM can be used as
270the target of optimizations: for example, through pointer analysis, it
271can be proven that a C automatic variable is never accessed outside of
272the current function... allowing it to be promoted to a simple SSA
273value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000274
Misha Brukman9d0919f2003-11-08 01:05:38 +0000275</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000276
Chris Lattner00950542001-06-06 20:29:01 +0000277<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000278<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000279
Misha Brukman9d0919f2003-11-08 01:05:38 +0000280<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000281
Chris Lattner261efe92003-11-25 01:02:51 +0000282<p>It is important to note that this document describes 'well formed'
283LLVM assembly language. There is a difference between what the parser
284accepts and what is considered 'well formed'. For example, the
285following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000286
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000287<div class="doc_code">
Chris Lattnerd7923912004-05-23 21:06:01 +0000288<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000289%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000290</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000291</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000292
Chris Lattner261efe92003-11-25 01:02:51 +0000293<p>...because the definition of <tt>%x</tt> does not dominate all of
294its uses. The LLVM infrastructure provides a verification pass that may
295be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000296automatically run by the parser after parsing input assembly and by
Gabor Greif04367bf2007-07-06 22:07:22 +0000297the optimizer before it outputs bitcode. The violations pointed out
Chris Lattner261efe92003-11-25 01:02:51 +0000298by the verifier pass indicate bugs in transformation passes or input to
299the parser.</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000300</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000301
Chris Lattnercc689392007-10-03 17:34:29 +0000302<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000303
Chris Lattner00950542001-06-06 20:29:01 +0000304<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000305<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000306<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000307
Misha Brukman9d0919f2003-11-08 01:05:38 +0000308<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000309
Reid Spencer2c452282007-08-07 14:34:28 +0000310 <p>LLVM identifiers come in two basic types: global and local. Global
311 identifiers (functions, global variables) begin with the @ character. Local
312 identifiers (register names, types) begin with the % character. Additionally,
313 there are three different formats for identifiers, for different purposes:
Chris Lattnerd7923912004-05-23 21:06:01 +0000314
Chris Lattner00950542001-06-06 20:29:01 +0000315<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000316 <li>Named values are represented as a string of characters with their prefix.
317 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
318 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Chris Lattnere5d947b2004-12-09 16:36:40 +0000319 Identifiers which require other characters in their names can be surrounded
Reid Spencer2c452282007-08-07 14:34:28 +0000320 with quotes. In this way, anything except a <tt>&quot;</tt> character can
321 be used in a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000322
Reid Spencer2c452282007-08-07 14:34:28 +0000323 <li>Unnamed values are represented as an unsigned numeric value with their
324 prefix. For example, %12, @2, %44.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000325
Reid Spencercc16dc32004-12-09 18:02:53 +0000326 <li>Constants, which are described in a <a href="#constants">section about
327 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000328</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000329
Reid Spencer2c452282007-08-07 14:34:28 +0000330<p>LLVM requires that values start with a prefix for two reasons: Compilers
Chris Lattnere5d947b2004-12-09 16:36:40 +0000331don't need to worry about name clashes with reserved words, and the set of
332reserved words may be expanded in the future without penalty. Additionally,
333unnamed identifiers allow a compiler to quickly come up with a temporary
334variable without having to avoid symbol table conflicts.</p>
335
Chris Lattner261efe92003-11-25 01:02:51 +0000336<p>Reserved words in LLVM are very similar to reserved words in other
Reid Spencer5c0ef472006-11-11 23:08:07 +0000337languages. There are keywords for different opcodes
338('<tt><a href="#i_add">add</a></tt>',
339 '<tt><a href="#i_bitcast">bitcast</a></tt>',
340 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
Reid Spencerca86e162006-12-31 07:07:53 +0000341href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
Chris Lattnere5d947b2004-12-09 16:36:40 +0000342and others. These reserved words cannot conflict with variable names, because
Reid Spencer2c452282007-08-07 14:34:28 +0000343none of them start with a prefix character ('%' or '@').</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000344
345<p>Here is an example of LLVM code to multiply the integer variable
346'<tt>%X</tt>' by 8:</p>
347
Misha Brukman9d0919f2003-11-08 01:05:38 +0000348<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000349
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000350<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000351<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000352%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000353</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000354</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000355
Misha Brukman9d0919f2003-11-08 01:05:38 +0000356<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000357
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000358<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000359<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000360%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000361</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000362</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000363
Misha Brukman9d0919f2003-11-08 01:05:38 +0000364<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000365
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000366<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000367<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000368<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
369<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
370%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000371</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000372</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000373
Chris Lattner261efe92003-11-25 01:02:51 +0000374<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
375important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000376
Chris Lattner00950542001-06-06 20:29:01 +0000377<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000378
379 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
380 line.</li>
381
382 <li>Unnamed temporaries are created when the result of a computation is not
383 assigned to a named value.</li>
384
Misha Brukman9d0919f2003-11-08 01:05:38 +0000385 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000386
Misha Brukman9d0919f2003-11-08 01:05:38 +0000387</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000388
John Criswelle4c57cc2005-05-12 16:52:32 +0000389<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000390demonstrating instructions, we will follow an instruction with a comment that
391defines the type and name of value produced. Comments are shown in italic
392text.</p>
393
Misha Brukman9d0919f2003-11-08 01:05:38 +0000394</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000395
396<!-- *********************************************************************** -->
397<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
398<!-- *********************************************************************** -->
399
400<!-- ======================================================================= -->
401<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
402</div>
403
404<div class="doc_text">
405
406<p>LLVM programs are composed of "Module"s, each of which is a
407translation unit of the input programs. Each module consists of
408functions, global variables, and symbol table entries. Modules may be
409combined together with the LLVM linker, which merges function (and
410global variable) definitions, resolves forward declarations, and merges
411symbol table entries. Here is an example of the "hello world" module:</p>
412
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000413<div class="doc_code">
Chris Lattnerfa730212004-12-09 16:11:40 +0000414<pre><i>; Declare the string constant as a global constant...</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000415<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
416 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000417
418<i>; External declaration of the puts function</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000419<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000420
421<i>; Definition of main function</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000422define i32 @main() { <i>; i32()* </i>
Reid Spencerca86e162006-12-31 07:07:53 +0000423 <i>; Convert [13x i8 ]* to i8 *...</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000424 %cast210 = <a
Chris Lattner6c0955b2007-06-12 17:01:15 +0000425 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000426
427 <i>; Call puts function to write out the string to stdout...</i>
428 <a
Chris Lattnera89e5f12007-06-12 17:00:26 +0000429 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000430 <a
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000431 href="#i_ret">ret</a> i32 0<br>}<br>
432</pre>
433</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000434
435<p>This example is made up of a <a href="#globalvars">global variable</a>
436named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
437function, and a <a href="#functionstructure">function definition</a>
438for "<tt>main</tt>".</p>
439
Chris Lattnere5d947b2004-12-09 16:36:40 +0000440<p>In general, a module is made up of a list of global values,
441where both functions and global variables are global values. Global values are
442represented by a pointer to a memory location (in this case, a pointer to an
443array of char, and a pointer to a function), and have one of the following <a
444href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000445
Chris Lattnere5d947b2004-12-09 16:36:40 +0000446</div>
447
448<!-- ======================================================================= -->
449<div class="doc_subsection">
450 <a name="linkage">Linkage Types</a>
451</div>
452
453<div class="doc_text">
454
455<p>
456All Global Variables and Functions have one of the following types of linkage:
457</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000458
459<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000460
Chris Lattnerfa730212004-12-09 16:11:40 +0000461 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000462
463 <dd>Global values with internal linkage are only directly accessible by
464 objects in the current module. In particular, linking code into a module with
465 an internal global value may cause the internal to be renamed as necessary to
466 avoid collisions. Because the symbol is internal to the module, all
467 references can be updated. This corresponds to the notion of the
Chris Lattner4887bd82007-01-14 06:51:48 +0000468 '<tt>static</tt>' keyword in C.
Chris Lattnerfa730212004-12-09 16:11:40 +0000469 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000470
Chris Lattnerfa730212004-12-09 16:11:40 +0000471 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000472
Chris Lattner4887bd82007-01-14 06:51:48 +0000473 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
474 the same name when linkage occurs. This is typically used to implement
475 inline functions, templates, or other code which must be generated in each
476 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
477 allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000478 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000479
Chris Lattnerfa730212004-12-09 16:11:40 +0000480 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000481
482 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
483 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
Chris Lattner4887bd82007-01-14 06:51:48 +0000484 used for globals that may be emitted in multiple translation units, but that
485 are not guaranteed to be emitted into every translation unit that uses them.
486 One example of this are common globals in C, such as "<tt>int X;</tt>" at
487 global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000488 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000489
Chris Lattnerfa730212004-12-09 16:11:40 +0000490 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000491
492 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
493 pointer to array type. When two global variables with appending linkage are
494 linked together, the two global arrays are appended together. This is the
495 LLVM, typesafe, equivalent of having the system linker append together
496 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000497 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000498
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000499 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
500 <dd>The semantics of this linkage follow the ELF model: the symbol is weak
501 until linked, if not linked, the symbol becomes null instead of being an
502 undefined reference.
503 </dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000504
Chris Lattnerfa730212004-12-09 16:11:40 +0000505 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000506
507 <dd>If none of the above identifiers are used, the global is externally
508 visible, meaning that it participates in linkage and can be used to resolve
509 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000510 </dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000511</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000512
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000513 <p>
514 The next two types of linkage are targeted for Microsoft Windows platform
515 only. They are designed to support importing (exporting) symbols from (to)
516 DLLs.
517 </p>
518
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000519 <dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000520 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
521
522 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
523 or variable via a global pointer to a pointer that is set up by the DLL
524 exporting the symbol. On Microsoft Windows targets, the pointer name is
525 formed by combining <code>_imp__</code> and the function or variable name.
526 </dd>
527
528 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
529
530 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
531 pointer to a pointer in a DLL, so that it can be referenced with the
532 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
533 name is formed by combining <code>_imp__</code> and the function or variable
534 name.
535 </dd>
536
Chris Lattnerfa730212004-12-09 16:11:40 +0000537</dl>
538
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000539<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
Chris Lattnerfa730212004-12-09 16:11:40 +0000540variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
541variable and was linked with this one, one of the two would be renamed,
542preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
543external (i.e., lacking any linkage declarations), they are accessible
Reid Spencerac8d2762007-01-05 00:59:10 +0000544outside of the current module.</p>
545<p>It is illegal for a function <i>declaration</i>
546to have any linkage type other than "externally visible", <tt>dllimport</tt>,
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000547or <tt>extern_weak</tt>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000548<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
549linkages.
Chris Lattnerfa730212004-12-09 16:11:40 +0000550</div>
551
552<!-- ======================================================================= -->
553<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000554 <a name="callingconv">Calling Conventions</a>
555</div>
556
557<div class="doc_text">
558
559<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
560and <a href="#i_invoke">invokes</a> can all have an optional calling convention
561specified for the call. The calling convention of any pair of dynamic
562caller/callee must match, or the behavior of the program is undefined. The
563following calling conventions are supported by LLVM, and more may be added in
564the future:</p>
565
566<dl>
567 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
568
569 <dd>This calling convention (the default if no other calling convention is
570 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000571 supports varargs function calls and tolerates some mismatch in the declared
Reid Spencerc28d2bc2006-12-31 21:30:18 +0000572 prototype and implemented declaration of the function (as does normal C).
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000573 </dd>
574
575 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
576
577 <dd>This calling convention attempts to make calls as fast as possible
578 (e.g. by passing things in registers). This calling convention allows the
579 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000580 without having to conform to an externally specified ABI. Implementations of
581 this convention should allow arbitrary tail call optimization to be supported.
582 This calling convention does not support varargs and requires the prototype of
583 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000584 </dd>
585
586 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
587
588 <dd>This calling convention attempts to make code in the caller as efficient
589 as possible under the assumption that the call is not commonly executed. As
590 such, these calls often preserve all registers so that the call does not break
591 any live ranges in the caller side. This calling convention does not support
592 varargs and requires the prototype of all callees to exactly match the
593 prototype of the function definition.
594 </dd>
595
Chris Lattnercfe6b372005-05-07 01:46:40 +0000596 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000597
598 <dd>Any calling convention may be specified by number, allowing
599 target-specific calling conventions to be used. Target specific calling
600 conventions start at 64.
601 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000602</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000603
604<p>More calling conventions can be added/defined on an as-needed basis, to
605support pascal conventions or any other well-known target-independent
606convention.</p>
607
608</div>
609
610<!-- ======================================================================= -->
611<div class="doc_subsection">
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000612 <a name="visibility">Visibility Styles</a>
613</div>
614
615<div class="doc_text">
616
617<p>
618All Global Variables and Functions have one of the following visibility styles:
619</p>
620
621<dl>
622 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
623
624 <dd>On ELF, default visibility means that the declaration is visible to other
625 modules and, in shared libraries, means that the declared entity may be
626 overridden. On Darwin, default visibility means that the declaration is
627 visible to other modules. Default visibility corresponds to "external
628 linkage" in the language.
629 </dd>
630
631 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
632
633 <dd>Two declarations of an object with hidden visibility refer to the same
634 object if they are in the same shared object. Usually, hidden visibility
635 indicates that the symbol will not be placed into the dynamic symbol table,
636 so no other module (executable or shared library) can reference it
637 directly.
638 </dd>
639
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000640 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
641
642 <dd>On ELF, protected visibility indicates that the symbol will be placed in
643 the dynamic symbol table, but that references within the defining module will
644 bind to the local symbol. That is, the symbol cannot be overridden by another
645 module.
646 </dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000647</dl>
648
649</div>
650
651<!-- ======================================================================= -->
652<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000653 <a name="globalvars">Global Variables</a>
654</div>
655
656<div class="doc_text">
657
Chris Lattner3689a342005-02-12 19:30:21 +0000658<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000659instead of run-time. Global variables may optionally be initialized, may have
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000660an explicit section to be placed in, and may have an optional explicit alignment
661specified. A variable may be defined as "thread_local", which means that it
662will not be shared by threads (each thread will have a separated copy of the
663variable). A variable may be defined as a global "constant," which indicates
664that the contents of the variable will <b>never</b> be modified (enabling better
Chris Lattner3689a342005-02-12 19:30:21 +0000665optimization, allowing the global data to be placed in the read-only section of
666an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000667cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000668
669<p>
670LLVM explicitly allows <em>declarations</em> of global variables to be marked
671constant, even if the final definition of the global is not. This capability
672can be used to enable slightly better optimization of the program, but requires
673the language definition to guarantee that optimizations based on the
674'constantness' are valid for the translation units that do not include the
675definition.
676</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000677
678<p>As SSA values, global variables define pointer values that are in
679scope (i.e. they dominate) all basic blocks in the program. Global
680variables always define a pointer to their "content" type because they
681describe a region of memory, and all memory objects in LLVM are
682accessed through pointers.</p>
683
Christopher Lamb284d9922007-12-11 09:31:00 +0000684<p>A global variable may be declared to reside in a target-specifc numbered
685address space. For targets that support them, address spaces may affect how
686optimizations are performed and/or what target instructions are used to access
Christopher Lambd49e18d2007-12-12 08:44:39 +0000687the variable. The default address space is zero. The address space qualifier
688must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000689
Chris Lattner88f6c462005-11-12 00:45:07 +0000690<p>LLVM allows an explicit section to be specified for globals. If the target
691supports it, it will emit globals to the section specified.</p>
692
Chris Lattner2cbdc452005-11-06 08:02:57 +0000693<p>An explicit alignment may be specified for a global. If not present, or if
694the alignment is set to zero, the alignment of the global is set by the target
695to whatever it feels convenient. If an explicit alignment is specified, the
696global is forced to have at least that much alignment. All alignments must be
697a power of 2.</p>
698
Christopher Lamb284d9922007-12-11 09:31:00 +0000699<p>For example, the following defines a global in a numbered address space with
700an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000701
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000702<div class="doc_code">
Chris Lattner68027ea2007-01-14 00:27:09 +0000703<pre>
Christopher Lamb284d9922007-12-11 09:31:00 +0000704@G = constant float 1.0 addrspace(5), section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000705</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000706</div>
Chris Lattner68027ea2007-01-14 00:27:09 +0000707
Chris Lattnerfa730212004-12-09 16:11:40 +0000708</div>
709
710
711<!-- ======================================================================= -->
712<div class="doc_subsection">
713 <a name="functionstructure">Functions</a>
714</div>
715
716<div class="doc_text">
717
Reid Spencerca86e162006-12-31 07:07:53 +0000718<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
719an optional <a href="#linkage">linkage type</a>, an optional
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000720<a href="#visibility">visibility style</a>, an optional
Reid Spencerca86e162006-12-31 07:07:53 +0000721<a href="#callingconv">calling convention</a>, a return type, an optional
722<a href="#paramattrs">parameter attribute</a> for the return type, a function
723name, a (possibly empty) argument list (each with optional
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000724<a href="#paramattrs">parameter attributes</a>), an optional section, an
Gordon Henriksene754abe2007-12-10 03:30:21 +0000725optional alignment, an optional <a href="#gc">garbage collector name</a>, an
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000726opening curly brace, a list of basic blocks, and a closing curly brace.
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000727
728LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
729optional <a href="#linkage">linkage type</a>, an optional
730<a href="#visibility">visibility style</a>, an optional
731<a href="#callingconv">calling convention</a>, a return type, an optional
Reid Spencerca86e162006-12-31 07:07:53 +0000732<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000733name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksene754abe2007-12-10 03:30:21 +0000734<a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000735
736<p>A function definition contains a list of basic blocks, forming the CFG for
737the function. Each basic block may optionally start with a label (giving the
738basic block a symbol table entry), contains a list of instructions, and ends
739with a <a href="#terminators">terminator</a> instruction (such as a branch or
740function return).</p>
741
Chris Lattner4a3c9012007-06-08 16:52:14 +0000742<p>The first basic block in a function is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000743executed on entrance to the function, and it is not allowed to have predecessor
744basic blocks (i.e. there can not be any branches to the entry block of a
745function). Because the block can have no predecessors, it also cannot have any
746<a href="#i_phi">PHI nodes</a>.</p>
747
Chris Lattner88f6c462005-11-12 00:45:07 +0000748<p>LLVM allows an explicit section to be specified for functions. If the target
749supports it, it will emit functions to the section specified.</p>
750
Chris Lattner2cbdc452005-11-06 08:02:57 +0000751<p>An explicit alignment may be specified for a function. If not present, or if
752the alignment is set to zero, the alignment of the function is set by the target
753to whatever it feels convenient. If an explicit alignment is specified, the
754function is forced to have at least that much alignment. All alignments must be
755a power of 2.</p>
756
Chris Lattnerfa730212004-12-09 16:11:40 +0000757</div>
758
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000759
760<!-- ======================================================================= -->
761<div class="doc_subsection">
762 <a name="aliasstructure">Aliases</a>
763</div>
764<div class="doc_text">
765 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov726d45c2008-03-22 08:36:14 +0000766 function, global variable, another alias or bitcast of global value). Aliases
767 may have an optional <a href="#linkage">linkage type</a>, and an
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000768 optional <a href="#visibility">visibility style</a>.</p>
769
770 <h5>Syntax:</h5>
771
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000772<div class="doc_code">
Bill Wendlingaac388b2007-05-29 09:42:13 +0000773<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000774@&lt;Name&gt; = [Linkage] [Visibility] alias &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000775</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000776</div>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000777
778</div>
779
780
781
Chris Lattner4e9aba72006-01-23 23:23:47 +0000782<!-- ======================================================================= -->
Reid Spencerca86e162006-12-31 07:07:53 +0000783<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
784<div class="doc_text">
785 <p>The return type and each parameter of a function type may have a set of
786 <i>parameter attributes</i> associated with them. Parameter attributes are
787 used to communicate additional information about the result or parameters of
Duncan Sandsdc024672007-11-27 13:23:08 +0000788 a function. Parameter attributes are considered to be part of the function,
789 not of the function type, so functions with different parameter attributes
790 can have the same function type.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000791
Reid Spencer950e9f82007-01-15 18:27:39 +0000792 <p>Parameter attributes are simple keywords that follow the type specified. If
793 multiple parameter attributes are needed, they are space separated. For
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000794 example:</p>
795
796<div class="doc_code">
797<pre>
Duncan Sandsdc024672007-11-27 13:23:08 +0000798declare i32 @printf(i8* noalias , ...) nounwind
799declare i32 @atoi(i8*) nounwind readonly
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000800</pre>
801</div>
802
Duncan Sandsdc024672007-11-27 13:23:08 +0000803 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
804 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000805
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000806 <p>Currently, only the following parameter attributes are defined:</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000807 <dl>
Reid Spencer9445e9a2007-07-19 23:13:04 +0000808 <dt><tt>zeroext</tt></dt>
Reid Spencerca86e162006-12-31 07:07:53 +0000809 <dd>This indicates that the parameter should be zero extended just before
810 a call to this function.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000811
Reid Spencer9445e9a2007-07-19 23:13:04 +0000812 <dt><tt>signext</tt></dt>
Reid Spencerca86e162006-12-31 07:07:53 +0000813 <dd>This indicates that the parameter should be sign extended just before
814 a call to this function.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000815
Anton Korobeynikov34d6dea2007-01-28 14:30:45 +0000816 <dt><tt>inreg</tt></dt>
817 <dd>This indicates that the parameter should be placed in register (if
Anton Korobeynikov66a8c8c2007-01-28 15:27:21 +0000818 possible) during assembling function call. Support for this attribute is
819 target-specific</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000820
821 <dt><tt>byval</tt></dt>
Chris Lattner0747baa2008-01-15 04:34:22 +0000822 <dd>This indicates that the pointer parameter should really be passed by
823 value to the function. The attribute implies that a hidden copy of the
824 pointee is made between the caller and the callee, so the callee is unable
825 to modify the value in the callee. This attribute is only valid on llvm
826 pointer arguments. It is generally used to pass structs and arrays by
827 value, but is also valid on scalars (even though this is silly).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000828
Anton Korobeynikov34d6dea2007-01-28 14:30:45 +0000829 <dt><tt>sret</tt></dt>
Duncan Sandse26dec62008-02-18 04:19:38 +0000830 <dd>This indicates that the pointer parameter specifies the address of a
831 structure that is the return value of the function in the source program.
Duncan Sandsb7f0c082008-03-17 12:17:41 +0000832 Loads and stores to the structure are assumed not to trap.
Duncan Sandse26dec62008-02-18 04:19:38 +0000833 May only be applied to the first parameter.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000834
Zhou Shengfebca342007-06-05 05:28:26 +0000835 <dt><tt>noalias</tt></dt>
Owen Anderson117bbd32008-02-18 04:09:01 +0000836 <dd>This indicates that the parameter does not alias any global or any other
837 parameter. The caller is responsible for ensuring that this is the case,
838 usually by placing the value in a stack allocation.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000839
Reid Spencer2dc52012007-03-22 02:18:56 +0000840 <dt><tt>noreturn</tt></dt>
841 <dd>This function attribute indicates that the function never returns. This
842 indicates to LLVM that every call to this function should be treated as if
843 an <tt>unreachable</tt> instruction immediately followed the call.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000844
Reid Spencer67606122007-03-22 02:02:11 +0000845 <dt><tt>nounwind</tt></dt>
Duncan Sandsb7f0c082008-03-17 12:17:41 +0000846 <dd>This function attribute indicates that no exceptions unwind out of the
847 function. Usually this is because the function makes no use of exceptions,
848 but it may also be that the function catches any exceptions thrown when
849 executing it.</dd>
850
Duncan Sands50f19f52007-07-27 19:57:41 +0000851 <dt><tt>nest</tt></dt>
852 <dd>This indicates that the parameter can be excised using the
853 <a href="#int_trampoline">trampoline intrinsics</a>.</dd>
Duncan Sandsed4a2f12007-11-22 20:23:04 +0000854 <dt><tt>readonly</tt></dt>
Duncan Sandsf04d5842007-11-14 21:14:02 +0000855 <dd>This function attribute indicates that the function has no side-effects
Duncan Sandsed4a2f12007-11-22 20:23:04 +0000856 except for producing a return value or throwing an exception. The value
857 returned must only depend on the function arguments and/or global variables.
858 It may use values obtained by dereferencing pointers.</dd>
859 <dt><tt>readnone</tt></dt>
860 <dd>A <tt>readnone</tt> function has the same restrictions as a <tt>readonly</tt>
Duncan Sandsf04d5842007-11-14 21:14:02 +0000861 function, but in addition it is not allowed to dereference any pointer arguments
862 or global variables.
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000863 </dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000864
Reid Spencerca86e162006-12-31 07:07:53 +0000865</div>
866
867<!-- ======================================================================= -->
Chris Lattner4e9aba72006-01-23 23:23:47 +0000868<div class="doc_subsection">
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000869 <a name="gc">Garbage Collector Names</a>
870</div>
871
872<div class="doc_text">
873<p>Each function may specify a garbage collector name, which is simply a
874string.</p>
875
876<div class="doc_code"><pre
877>define void @f() gc "name" { ...</pre></div>
878
879<p>The compiler declares the supported values of <i>name</i>. Specifying a
880collector which will cause the compiler to alter its output in order to support
881the named garbage collection algorithm.</p>
882</div>
883
884<!-- ======================================================================= -->
885<div class="doc_subsection">
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000886 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000887</div>
888
889<div class="doc_text">
890<p>
891Modules may contain "module-level inline asm" blocks, which corresponds to the
892GCC "file scope inline asm" blocks. These blocks are internally concatenated by
893LLVM and treated as a single unit, but may be separated in the .ll file if
894desired. The syntax is very simple:
895</p>
896
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000897<div class="doc_code">
898<pre>
899module asm "inline asm code goes here"
900module asm "more can go here"
901</pre>
902</div>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000903
904<p>The strings can contain any character by escaping non-printable characters.
905 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
906 for the number.
907</p>
908
909<p>
910 The inline asm code is simply printed to the machine code .s file when
911 assembly code is generated.
912</p>
913</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000914
Reid Spencerde151942007-02-19 23:54:10 +0000915<!-- ======================================================================= -->
916<div class="doc_subsection">
917 <a name="datalayout">Data Layout</a>
918</div>
919
920<div class="doc_text">
921<p>A module may specify a target specific data layout string that specifies how
Reid Spencerc8910842007-04-11 23:49:50 +0000922data is to be laid out in memory. The syntax for the data layout is simply:</p>
923<pre> target datalayout = "<i>layout specification</i>"</pre>
924<p>The <i>layout specification</i> consists of a list of specifications
925separated by the minus sign character ('-'). Each specification starts with a
926letter and may include other information after the letter to define some
927aspect of the data layout. The specifications accepted are as follows: </p>
Reid Spencerde151942007-02-19 23:54:10 +0000928<dl>
929 <dt><tt>E</tt></dt>
930 <dd>Specifies that the target lays out data in big-endian form. That is, the
931 bits with the most significance have the lowest address location.</dd>
932 <dt><tt>e</tt></dt>
933 <dd>Specifies that hte target lays out data in little-endian form. That is,
934 the bits with the least significance have the lowest address location.</dd>
935 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
936 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
937 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
938 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
939 too.</dd>
940 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
941 <dd>This specifies the alignment for an integer type of a given bit
942 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
943 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
944 <dd>This specifies the alignment for a vector type of a given bit
945 <i>size</i>.</dd>
946 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
947 <dd>This specifies the alignment for a floating point type of a given bit
948 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
949 (double).</dd>
950 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
951 <dd>This specifies the alignment for an aggregate type of a given bit
952 <i>size</i>.</dd>
953</dl>
954<p>When constructing the data layout for a given target, LLVM starts with a
955default set of specifications which are then (possibly) overriden by the
956specifications in the <tt>datalayout</tt> keyword. The default specifications
957are given in this list:</p>
958<ul>
959 <li><tt>E</tt> - big endian</li>
960 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
961 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
962 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
963 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
964 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
965 <li><tt>i64:32:64</tt> - i64 has abi alignment of 32-bits but preferred
966 alignment of 64-bits</li>
967 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
968 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
969 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
970 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
971 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
972</ul>
973<p>When llvm is determining the alignment for a given type, it uses the
974following rules:
975<ol>
976 <li>If the type sought is an exact match for one of the specifications, that
977 specification is used.</li>
978 <li>If no match is found, and the type sought is an integer type, then the
979 smallest integer type that is larger than the bitwidth of the sought type is
980 used. If none of the specifications are larger than the bitwidth then the the
981 largest integer type is used. For example, given the default specifications
982 above, the i7 type will use the alignment of i8 (next largest) while both
983 i65 and i256 will use the alignment of i64 (largest specified).</li>
984 <li>If no match is found, and the type sought is a vector type, then the
985 largest vector type that is smaller than the sought vector type will be used
986 as a fall back. This happens because <128 x double> can be implemented in
987 terms of 64 <2 x double>, for example.</li>
988</ol>
989</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000990
Chris Lattner00950542001-06-06 20:29:01 +0000991<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000992<div class="doc_section"> <a name="typesystem">Type System</a> </div>
993<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000994
Misha Brukman9d0919f2003-11-08 01:05:38 +0000995<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000996
Misha Brukman9d0919f2003-11-08 01:05:38 +0000997<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000998intermediate representation. Being typed enables a number of
999optimizations to be performed on the IR directly, without having to do
1000extra analyses on the side before the transformation. A strong type
1001system makes it easier to read the generated code and enables novel
1002analyses and transformations that are not feasible to perform on normal
1003three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +00001004
1005</div>
1006
Chris Lattner00950542001-06-06 20:29:01 +00001007<!-- ======================================================================= -->
Chris Lattner4f69f462008-01-04 04:32:38 +00001008<div class="doc_subsection"> <a name="t_classifications">Type
Chris Lattner261efe92003-11-25 01:02:51 +00001009Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001010<div class="doc_text">
Chris Lattner4f69f462008-01-04 04:32:38 +00001011<p>The types fall into a few useful
Chris Lattner261efe92003-11-25 01:02:51 +00001012classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001013
1014<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001015 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001016 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +00001017 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001018 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +00001019 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +00001020 </tr>
1021 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001022 <td><a href="#t_floating">floating point</a></td>
1023 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +00001024 </tr>
1025 <tr>
1026 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +00001027 <td><a href="#t_integer">integer</a>,
1028 <a href="#t_floating">floating point</a>,
1029 <a href="#t_pointer">pointer</a>,
1030 <a href="#t_vector">vector</a>
Reid Spencerca86e162006-12-31 07:07:53 +00001031 </td>
Chris Lattner261efe92003-11-25 01:02:51 +00001032 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001033 <tr>
1034 <td><a href="#t_primitive">primitive</a></td>
1035 <td><a href="#t_label">label</a>,
1036 <a href="#t_void">void</a>,
1037 <a href="#t_integer">integer</a>,
1038 <a href="#t_floating">floating point</a>.</td>
1039 </tr>
1040 <tr>
1041 <td><a href="#t_derived">derived</a></td>
1042 <td><a href="#t_integer">integer</a>,
1043 <a href="#t_array">array</a>,
1044 <a href="#t_function">function</a>,
1045 <a href="#t_pointer">pointer</a>,
1046 <a href="#t_struct">structure</a>,
1047 <a href="#t_pstruct">packed structure</a>,
1048 <a href="#t_vector">vector</a>,
1049 <a href="#t_opaque">opaque</a>.
1050 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +00001051 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001052</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001053
Chris Lattner261efe92003-11-25 01:02:51 +00001054<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1055most important. Values of these types are the only ones which can be
1056produced by instructions, passed as arguments, or used as operands to
1057instructions. This means that all structures and arrays must be
1058manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001059</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001060
Chris Lattner00950542001-06-06 20:29:01 +00001061<!-- ======================================================================= -->
Chris Lattner4f69f462008-01-04 04:32:38 +00001062<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner8f8c7b72008-01-04 04:34:14 +00001063
Chris Lattner4f69f462008-01-04 04:32:38 +00001064<div class="doc_text">
1065<p>The primitive types are the fundamental building blocks of the LLVM
1066system.</p>
1067
Chris Lattner8f8c7b72008-01-04 04:34:14 +00001068</div>
1069
Chris Lattner4f69f462008-01-04 04:32:38 +00001070<!-- _______________________________________________________________________ -->
1071<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1072
1073<div class="doc_text">
1074 <table>
1075 <tbody>
1076 <tr><th>Type</th><th>Description</th></tr>
1077 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1078 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1079 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1080 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1081 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1082 </tbody>
1083 </table>
1084</div>
1085
1086<!-- _______________________________________________________________________ -->
1087<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1088
1089<div class="doc_text">
1090<h5>Overview:</h5>
1091<p>The void type does not represent any value and has no size.</p>
1092
1093<h5>Syntax:</h5>
1094
1095<pre>
1096 void
1097</pre>
1098</div>
1099
1100<!-- _______________________________________________________________________ -->
1101<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1102
1103<div class="doc_text">
1104<h5>Overview:</h5>
1105<p>The label type represents code labels.</p>
1106
1107<h5>Syntax:</h5>
1108
1109<pre>
1110 label
1111</pre>
1112</div>
1113
1114
1115<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001116<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001117
Misha Brukman9d0919f2003-11-08 01:05:38 +00001118<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001119
Chris Lattner261efe92003-11-25 01:02:51 +00001120<p>The real power in LLVM comes from the derived types in the system.
1121This is what allows a programmer to represent arrays, functions,
1122pointers, and other useful types. Note that these derived types may be
1123recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001124
Misha Brukman9d0919f2003-11-08 01:05:38 +00001125</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001126
Chris Lattner00950542001-06-06 20:29:01 +00001127<!-- _______________________________________________________________________ -->
Reid Spencer2b916312007-05-16 18:44:01 +00001128<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1129
1130<div class="doc_text">
1131
1132<h5>Overview:</h5>
1133<p>The integer type is a very simple derived type that simply specifies an
1134arbitrary bit width for the integer type desired. Any bit width from 1 bit to
11352^23-1 (about 8 million) can be specified.</p>
1136
1137<h5>Syntax:</h5>
1138
1139<pre>
1140 iN
1141</pre>
1142
1143<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1144value.</p>
1145
1146<h5>Examples:</h5>
1147<table class="layout">
Chris Lattnerb9488a62007-12-18 06:18:21 +00001148 <tbody>
1149 <tr>
1150 <td><tt>i1</tt></td>
1151 <td>a single-bit integer.</td>
1152 </tr><tr>
1153 <td><tt>i32</tt></td>
1154 <td>a 32-bit integer.</td>
1155 </tr><tr>
1156 <td><tt>i1942652</tt></td>
1157 <td>a really big integer of over 1 million bits.</td>
Reid Spencer2b916312007-05-16 18:44:01 +00001158 </tr>
Chris Lattnerb9488a62007-12-18 06:18:21 +00001159 </tbody>
Reid Spencer2b916312007-05-16 18:44:01 +00001160</table>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001161</div>
Reid Spencer2b916312007-05-16 18:44:01 +00001162
1163<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001164<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001165
Misha Brukman9d0919f2003-11-08 01:05:38 +00001166<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001167
Chris Lattner00950542001-06-06 20:29:01 +00001168<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001169
Misha Brukman9d0919f2003-11-08 01:05:38 +00001170<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +00001171sequentially in memory. The array type requires a size (number of
1172elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001173
Chris Lattner7faa8832002-04-14 06:13:44 +00001174<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001175
1176<pre>
1177 [&lt;# elements&gt; x &lt;elementtype&gt;]
1178</pre>
1179
John Criswelle4c57cc2005-05-12 16:52:32 +00001180<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +00001181be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001182
Chris Lattner7faa8832002-04-14 06:13:44 +00001183<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001184<table class="layout">
1185 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001186 <td class="left"><tt>[40 x i32]</tt></td>
1187 <td class="left">Array of 40 32-bit integer values.</td>
1188 </tr>
1189 <tr class="layout">
1190 <td class="left"><tt>[41 x i32]</tt></td>
1191 <td class="left">Array of 41 32-bit integer values.</td>
1192 </tr>
1193 <tr class="layout">
1194 <td class="left"><tt>[4 x i8]</tt></td>
1195 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001196 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001197</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001198<p>Here are some examples of multidimensional arrays:</p>
1199<table class="layout">
1200 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001201 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1202 <td class="left">3x4 array of 32-bit integer values.</td>
1203 </tr>
1204 <tr class="layout">
1205 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1206 <td class="left">12x10 array of single precision floating point values.</td>
1207 </tr>
1208 <tr class="layout">
1209 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1210 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001211 </tr>
1212</table>
Chris Lattnere67a9512005-06-24 17:22:57 +00001213
John Criswell0ec250c2005-10-24 16:17:18 +00001214<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1215length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +00001216LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1217As a special case, however, zero length arrays are recognized to be variable
1218length. This allows implementation of 'pascal style arrays' with the LLVM
Reid Spencerca86e162006-12-31 07:07:53 +00001219type "{ i32, [0 x float]}", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +00001220
Misha Brukman9d0919f2003-11-08 01:05:38 +00001221</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001222
Chris Lattner00950542001-06-06 20:29:01 +00001223<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001224<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001225<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001226<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001227<p>The function type can be thought of as a function signature. It
Devang Patela582f402008-03-24 05:35:41 +00001228consists of a return type and a list of formal parameter types. The
Devang Patel7a3ad1a2008-03-24 20:52:42 +00001229return type of a function type is a scalar type or a void type or a struct type.
1230If the return type is a struct type then all struct elements must be of first
Devang Patel3a5bff82008-03-24 18:10:52 +00001231class types. Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +00001232(which are structures of pointers to functions), for indirect function
1233calls, and when defining a function.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001234
Chris Lattner00950542001-06-06 20:29:01 +00001235<h5>Syntax:</h5>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001236<pre> &lt;returntype list&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +00001237<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +00001238specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +00001239which indicates that the function takes a variable number of arguments.
1240Variable argument functions can access their arguments with the <a
Devang Patelc3fc6df2008-03-10 20:49:15 +00001241 href="#int_varargs">variable argument handling intrinsic</a> functions.
1242'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1243<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001244<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001245<table class="layout">
1246 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +00001247 <td class="left"><tt>i32 (i32)</tt></td>
1248 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001249 </td>
Reid Spencer92f82302006-12-31 07:18:34 +00001250 </tr><tr class="layout">
Reid Spencer9445e9a2007-07-19 23:13:04 +00001251 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +00001252 </tt></td>
Reid Spencer92f82302006-12-31 07:18:34 +00001253 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1254 an <tt>i16</tt> that should be sign extended and a
Reid Spencerca86e162006-12-31 07:07:53 +00001255 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
Reid Spencer92f82302006-12-31 07:18:34 +00001256 <tt>float</tt>.
1257 </td>
1258 </tr><tr class="layout">
1259 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1260 <td class="left">A vararg function that takes at least one
Reid Spencera5173382007-01-04 16:43:23 +00001261 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
Reid Spencer92f82302006-12-31 07:18:34 +00001262 which returns an integer. This is the signature for <tt>printf</tt> in
1263 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +00001264 </td>
Devang Patela582f402008-03-24 05:35:41 +00001265 </tr><tr class="layout">
1266 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Devang Patel3a5bff82008-03-24 18:10:52 +00001267 <td class="left">A function taking an <tt>i32></tt>, returning two
1268 <tt> i32 </tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Patela582f402008-03-24 05:35:41 +00001269 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001270 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001271</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001272
Misha Brukman9d0919f2003-11-08 01:05:38 +00001273</div>
Chris Lattner00950542001-06-06 20:29:01 +00001274<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001275<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001276<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001277<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001278<p>The structure type is used to represent a collection of data members
1279together in memory. The packing of the field types is defined to match
1280the ABI of the underlying processor. The elements of a structure may
1281be any type that has a size.</p>
1282<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1283and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1284field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1285instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001286<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001287<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +00001288<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001289<table class="layout">
1290 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001291 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1292 <td class="left">A triple of three <tt>i32</tt> values</td>
1293 </tr><tr class="layout">
1294 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1295 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1296 second element is a <a href="#t_pointer">pointer</a> to a
1297 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1298 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001299 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001300</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001301</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001302
Chris Lattner00950542001-06-06 20:29:01 +00001303<!-- _______________________________________________________________________ -->
Andrew Lenharth75e10682006-12-08 17:13:00 +00001304<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1305</div>
1306<div class="doc_text">
1307<h5>Overview:</h5>
1308<p>The packed structure type is used to represent a collection of data members
1309together in memory. There is no padding between fields. Further, the alignment
1310of a packed structure is 1 byte. The elements of a packed structure may
1311be any type that has a size.</p>
1312<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1313and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1314field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1315instruction.</p>
1316<h5>Syntax:</h5>
1317<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1318<h5>Examples:</h5>
1319<table class="layout">
1320 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001321 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1322 <td class="left">A triple of three <tt>i32</tt> values</td>
1323 </tr><tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001324 <td class="left"><tt>&lt; { float, i32 (i32)* } &gt;</tt></td>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001325 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1326 second element is a <a href="#t_pointer">pointer</a> to a
1327 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1328 an <tt>i32</tt>.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +00001329 </tr>
1330</table>
1331</div>
1332
1333<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001334<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001335<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +00001336<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001337<p>As in many languages, the pointer type represents a pointer or
Christopher Lamb284d9922007-12-11 09:31:00 +00001338reference to another object, which must live in memory. Pointer types may have
1339an optional address space attribute defining the target-specific numbered
1340address space where the pointed-to object resides. The default address space is
1341zero.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +00001342<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001343<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +00001344<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001345<table class="layout">
1346 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001347 <td class="left"><tt>[4x i32]*</tt></td>
1348 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1349 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1350 </tr>
1351 <tr class="layout">
1352 <td class="left"><tt>i32 (i32 *) *</tt></td>
1353 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +00001354 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +00001355 <tt>i32</tt>.</td>
1356 </tr>
1357 <tr class="layout">
1358 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1359 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1360 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001361 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001362</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001363</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001364
Chris Lattnera58561b2004-08-12 19:12:28 +00001365<!-- _______________________________________________________________________ -->
Reid Spencer485bad12007-02-15 03:07:05 +00001366<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001367<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +00001368
Chris Lattnera58561b2004-08-12 19:12:28 +00001369<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001370
Reid Spencer485bad12007-02-15 03:07:05 +00001371<p>A vector type is a simple derived type that represents a vector
1372of elements. Vector types are used when multiple primitive data
Chris Lattnera58561b2004-08-12 19:12:28 +00001373are operated in parallel using a single instruction (SIMD).
Reid Spencer485bad12007-02-15 03:07:05 +00001374A vector type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +00001375elements) and an underlying primitive data type. Vectors must have a power
Reid Spencer485bad12007-02-15 03:07:05 +00001376of two length (1, 2, 4, 8, 16 ...). Vector types are
Chris Lattnera58561b2004-08-12 19:12:28 +00001377considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001378
Chris Lattnera58561b2004-08-12 19:12:28 +00001379<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001380
1381<pre>
1382 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1383</pre>
1384
John Criswellc1f786c2005-05-13 22:25:59 +00001385<p>The number of elements is a constant integer value; elementtype may
Chris Lattner3b19d652007-01-15 01:54:13 +00001386be any integer or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001387
Chris Lattnera58561b2004-08-12 19:12:28 +00001388<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001389
Reid Spencerd3f876c2004-11-01 08:19:36 +00001390<table class="layout">
1391 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001392 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1393 <td class="left">Vector of 4 32-bit integer values.</td>
1394 </tr>
1395 <tr class="layout">
1396 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1397 <td class="left">Vector of 8 32-bit floating-point values.</td>
1398 </tr>
1399 <tr class="layout">
1400 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1401 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001402 </tr>
1403</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001404</div>
1405
Chris Lattner69c11bb2005-04-25 17:34:15 +00001406<!-- _______________________________________________________________________ -->
1407<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1408<div class="doc_text">
1409
1410<h5>Overview:</h5>
1411
1412<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksen8ac04ff2007-10-14 00:34:53 +00001413corresponds (for example) to the C notion of a forward declared structure type.
Chris Lattner69c11bb2005-04-25 17:34:15 +00001414In LLVM, opaque types can eventually be resolved to any type (not just a
1415structure type).</p>
1416
1417<h5>Syntax:</h5>
1418
1419<pre>
1420 opaque
1421</pre>
1422
1423<h5>Examples:</h5>
1424
1425<table class="layout">
1426 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001427 <td class="left"><tt>opaque</tt></td>
1428 <td class="left">An opaque type.</td>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001429 </tr>
1430</table>
1431</div>
1432
1433
Chris Lattnerc3f59762004-12-09 17:30:23 +00001434<!-- *********************************************************************** -->
1435<div class="doc_section"> <a name="constants">Constants</a> </div>
1436<!-- *********************************************************************** -->
1437
1438<div class="doc_text">
1439
1440<p>LLVM has several different basic types of constants. This section describes
1441them all and their syntax.</p>
1442
1443</div>
1444
1445<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +00001446<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001447
1448<div class="doc_text">
1449
1450<dl>
1451 <dt><b>Boolean constants</b></dt>
1452
1453 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Reid Spencerc78f3372007-01-12 03:35:51 +00001454 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
Chris Lattnerc3f59762004-12-09 17:30:23 +00001455 </dd>
1456
1457 <dt><b>Integer constants</b></dt>
1458
Reid Spencercc16dc32004-12-09 18:02:53 +00001459 <dd>Standard integers (such as '4') are constants of the <a
Reid Spencera5173382007-01-04 16:43:23 +00001460 href="#t_integer">integer</a> type. Negative numbers may be used with
Chris Lattnerc3f59762004-12-09 17:30:23 +00001461 integer types.
1462 </dd>
1463
1464 <dt><b>Floating point constants</b></dt>
1465
1466 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1467 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnera73afe02008-04-01 18:45:27 +00001468 notation (see below). The assembler requires the exact decimal value of
1469 a floating-point constant. For example, the assembler accepts 1.25 but
1470 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1471 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001472
1473 <dt><b>Null pointer constants</b></dt>
1474
John Criswell9e2485c2004-12-10 15:51:16 +00001475 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +00001476 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1477
1478</dl>
1479
John Criswell9e2485c2004-12-10 15:51:16 +00001480<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +00001481of floating point constants. For example, the form '<tt>double
14820x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
14834.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +00001484(and the only time that they are generated by the disassembler) is when a
1485floating point constant must be emitted but it cannot be represented as a
1486decimal floating point number. For example, NaN's, infinities, and other
1487special values are represented in their IEEE hexadecimal format so that
1488assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001489
1490</div>
1491
1492<!-- ======================================================================= -->
1493<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1494</div>
1495
1496<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001497<p>Aggregate constants arise from aggregation of simple constants
1498and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001499
1500<dl>
1501 <dt><b>Structure constants</b></dt>
1502
1503 <dd>Structure constants are represented with notation similar to structure
1504 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner64910ee2007-12-25 20:34:52 +00001505 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1506 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>". Structure constants
Chris Lattnerd4f6b172005-03-07 22:13:59 +00001507 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +00001508 types of elements must match those specified by the type.
1509 </dd>
1510
1511 <dt><b>Array constants</b></dt>
1512
1513 <dd>Array constants are represented with notation similar to array type
1514 definitions (a comma separated list of elements, surrounded by square brackets
Reid Spencerca86e162006-12-31 07:07:53 +00001515 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +00001516 constants must have <a href="#t_array">array type</a>, and the number and
1517 types of elements must match those specified by the type.
1518 </dd>
1519
Reid Spencer485bad12007-02-15 03:07:05 +00001520 <dt><b>Vector constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001521
Reid Spencer485bad12007-02-15 03:07:05 +00001522 <dd>Vector constants are represented with notation similar to vector type
Chris Lattnerc3f59762004-12-09 17:30:23 +00001523 definitions (a comma separated list of elements, surrounded by
Reid Spencerca86e162006-12-31 07:07:53 +00001524 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001525 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
Reid Spencer485bad12007-02-15 03:07:05 +00001526 href="#t_vector">vector type</a>, and the number and types of elements must
Chris Lattnerc3f59762004-12-09 17:30:23 +00001527 match those specified by the type.
1528 </dd>
1529
1530 <dt><b>Zero initialization</b></dt>
1531
1532 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1533 value to zero of <em>any</em> type, including scalar and aggregate types.
1534 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +00001535 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +00001536 initializers.
1537 </dd>
1538</dl>
1539
1540</div>
1541
1542<!-- ======================================================================= -->
1543<div class="doc_subsection">
1544 <a name="globalconstants">Global Variable and Function Addresses</a>
1545</div>
1546
1547<div class="doc_text">
1548
1549<p>The addresses of <a href="#globalvars">global variables</a> and <a
1550href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +00001551constants. These constants are explicitly referenced when the <a
1552href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +00001553href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1554file:</p>
1555
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001556<div class="doc_code">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001557<pre>
Chris Lattnera18a4242007-06-06 18:28:13 +00001558@X = global i32 17
1559@Y = global i32 42
1560@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +00001561</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001562</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001563
1564</div>
1565
1566<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +00001567<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001568<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +00001569 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +00001570 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +00001571 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001572
Reid Spencer2dc45b82004-12-09 18:13:12 +00001573 <p>Undefined values indicate to the compiler that the program is well defined
1574 no matter what value is used, giving the compiler more freedom to optimize.
1575 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001576</div>
1577
1578<!-- ======================================================================= -->
1579<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1580</div>
1581
1582<div class="doc_text">
1583
1584<p>Constant expressions are used to allow expressions involving other constants
1585to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +00001586href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +00001587that does not have side effects (e.g. load and call are not supported). The
1588following is the syntax for constant expressions:</p>
1589
1590<dl>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001591 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1592 <dd>Truncate a constant to another type. The bit size of CST must be larger
Chris Lattner3b19d652007-01-15 01:54:13 +00001593 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001594
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001595 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1596 <dd>Zero extend a constant to another type. The bit size of CST must be
Chris Lattner3b19d652007-01-15 01:54:13 +00001597 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001598
1599 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1600 <dd>Sign extend a constant to another type. The bit size of CST must be
Chris Lattner3b19d652007-01-15 01:54:13 +00001601 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001602
1603 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1604 <dd>Truncate a floating point constant to another floating point type. The
1605 size of CST must be larger than the size of TYPE. Both types must be
1606 floating point.</dd>
1607
1608 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1609 <dd>Floating point extend a constant to another type. The size of CST must be
1610 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1611
Reid Spencer1539a1c2007-07-31 14:40:14 +00001612 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001613 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begemanb348d182007-11-17 03:58:34 +00001614 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1615 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1616 of the same number of elements. If the value won't fit in the integer type,
1617 the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001618
Reid Spencerd4448792006-11-09 23:03:26 +00001619 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001620 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begemanb348d182007-11-17 03:58:34 +00001621 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1622 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1623 of the same number of elements. If the value won't fit in the integer type,
1624 the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001625
Reid Spencerd4448792006-11-09 23:03:26 +00001626 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001627 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begemanb348d182007-11-17 03:58:34 +00001628 constant. TYPE must be a scalar or vector floating point type. CST must be of
1629 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1630 of the same number of elements. If the value won't fit in the floating point
1631 type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001632
Reid Spencerd4448792006-11-09 23:03:26 +00001633 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001634 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begemanb348d182007-11-17 03:58:34 +00001635 constant. TYPE must be a scalar or vector floating point type. CST must be of
1636 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1637 of the same number of elements. If the value won't fit in the floating point
1638 type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001639
Reid Spencer5c0ef472006-11-11 23:08:07 +00001640 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1641 <dd>Convert a pointer typed constant to the corresponding integer constant
1642 TYPE must be an integer type. CST must be of pointer type. The CST value is
1643 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1644
1645 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1646 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1647 pointer type. CST must be of integer type. The CST value is zero extended,
1648 truncated, or unchanged to make it fit in a pointer size. This one is
1649 <i>really</i> dangerous!</dd>
1650
1651 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001652 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1653 identical (same number of bits). The conversion is done as if the CST value
1654 was stored to memory and read back as TYPE. In other words, no bits change
Reid Spencer5c0ef472006-11-11 23:08:07 +00001655 with this operator, just the type. This can be used for conversion of
Reid Spencer485bad12007-02-15 03:07:05 +00001656 vector types to any other type, as long as they have the same bit width. For
Reid Spencer5c0ef472006-11-11 23:08:07 +00001657 pointers it is only valid to cast to another pointer type.
Reid Spencer9dee3ac2006-11-08 01:11:31 +00001658 </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001659
1660 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1661
1662 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1663 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1664 instruction, the index list may have zero or more indexes, which are required
1665 to make sense for the type of "CSTPTR".</dd>
1666
Robert Bocchino9fbe1452006-01-10 19:31:34 +00001667 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1668
1669 <dd>Perform the <a href="#i_select">select operation</a> on
Reid Spencer01c42592006-12-04 19:23:19 +00001670 constants.</dd>
1671
1672 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1673 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1674
1675 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1676 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +00001677
1678 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1679
1680 <dd>Perform the <a href="#i_extractelement">extractelement
1681 operation</a> on constants.
1682
Robert Bocchino05ccd702006-01-15 20:48:27 +00001683 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1684
1685 <dd>Perform the <a href="#i_insertelement">insertelement
Reid Spencer01c42592006-12-04 19:23:19 +00001686 operation</a> on constants.</dd>
Robert Bocchino05ccd702006-01-15 20:48:27 +00001687
Chris Lattnerc1989542006-04-08 00:13:41 +00001688
1689 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1690
1691 <dd>Perform the <a href="#i_shufflevector">shufflevector
Reid Spencer01c42592006-12-04 19:23:19 +00001692 operation</a> on constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +00001693
Chris Lattnerc3f59762004-12-09 17:30:23 +00001694 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1695
Reid Spencer2dc45b82004-12-09 18:13:12 +00001696 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1697 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +00001698 binary</a> operations. The constraints on operands are the same as those for
1699 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +00001700 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001701</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001702</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00001703
Chris Lattner00950542001-06-06 20:29:01 +00001704<!-- *********************************************************************** -->
Chris Lattnere87d6532006-01-25 23:47:57 +00001705<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1706<!-- *********************************************************************** -->
1707
1708<!-- ======================================================================= -->
1709<div class="doc_subsection">
1710<a name="inlineasm">Inline Assembler Expressions</a>
1711</div>
1712
1713<div class="doc_text">
1714
1715<p>
1716LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1717Module-Level Inline Assembly</a>) through the use of a special value. This
1718value represents the inline assembler as a string (containing the instructions
1719to emit), a list of operand constraints (stored as a string), and a flag that
1720indicates whether or not the inline asm expression has side effects. An example
1721inline assembler expression is:
1722</p>
1723
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001724<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +00001725<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001726i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +00001727</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001728</div>
Chris Lattnere87d6532006-01-25 23:47:57 +00001729
1730<p>
1731Inline assembler expressions may <b>only</b> be used as the callee operand of
1732a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1733</p>
1734
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001735<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +00001736<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001737%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattnere87d6532006-01-25 23:47:57 +00001738</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001739</div>
Chris Lattnere87d6532006-01-25 23:47:57 +00001740
1741<p>
1742Inline asms with side effects not visible in the constraint list must be marked
1743as having side effects. This is done through the use of the
1744'<tt>sideeffect</tt>' keyword, like so:
1745</p>
1746
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001747<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +00001748<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001749call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +00001750</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001751</div>
Chris Lattnere87d6532006-01-25 23:47:57 +00001752
1753<p>TODO: The format of the asm and constraints string still need to be
1754documented here. Constraints on what can be done (e.g. duplication, moving, etc
1755need to be documented).
1756</p>
1757
1758</div>
1759
1760<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00001761<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1762<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00001763
Misha Brukman9d0919f2003-11-08 01:05:38 +00001764<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001765
Chris Lattner261efe92003-11-25 01:02:51 +00001766<p>The LLVM instruction set consists of several different
1767classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +00001768instructions</a>, <a href="#binaryops">binary instructions</a>,
1769<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +00001770 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1771instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001772
Misha Brukman9d0919f2003-11-08 01:05:38 +00001773</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001774
Chris Lattner00950542001-06-06 20:29:01 +00001775<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00001776<div class="doc_subsection"> <a name="terminators">Terminator
1777Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001778
Misha Brukman9d0919f2003-11-08 01:05:38 +00001779<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +00001780
Chris Lattner261efe92003-11-25 01:02:51 +00001781<p>As mentioned <a href="#functionstructure">previously</a>, every
1782basic block in a program ends with a "Terminator" instruction, which
1783indicates which block should be executed after the current block is
1784finished. These terminator instructions typically yield a '<tt>void</tt>'
1785value: they produce control flow, not values (the one exception being
1786the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +00001787<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +00001788 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1789instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +00001790the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1791 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1792 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001793
Misha Brukman9d0919f2003-11-08 01:05:38 +00001794</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001795
Chris Lattner00950542001-06-06 20:29:01 +00001796<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001797<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1798Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001799<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001800<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001801<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 +00001802 ret void <i>; Return from void function</i>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001803 ret &lt;type&gt; &lt;value&gt;, &lt;type&gt; &lt;value&gt; <i>; Return two values from a non-void function </i>
Chris Lattner00950542001-06-06 20:29:01 +00001804</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001805<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001806<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +00001807value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +00001808<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +00001809returns a value and then causes control flow, and one that just causes
1810control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001811<h5>Arguments:</h5>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001812<p>The '<tt>ret</tt>' instruction may return one or multiple values. The
Devang Patel0dbb4a12008-03-11 05:51:59 +00001813type of each return value must be a '<a href="#t_firstclass">first class</a>'
1814 type. Note that a function is not <a href="#wellformed">well formed</a>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001815if there exists a '<tt>ret</tt>' instruction inside of the function that
Devang Patel0dbb4a12008-03-11 05:51:59 +00001816returns values that do not match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001817<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001818<p>When the '<tt>ret</tt>' instruction is executed, control flow
1819returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +00001820 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +00001821the instruction after the call. If the caller was an "<a
1822 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +00001823at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +00001824returns a value, that value shall set the call or invoke instruction's
Devang Patelc3fc6df2008-03-10 20:49:15 +00001825return value. If the instruction returns multiple values then these
Devang Patel0dbb4a12008-03-11 05:51:59 +00001826values can only be accessed through a '<a href="#i_getresult"><tt>getresult</tt>
1827</a>' instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001828<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00001829<pre> ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00001830 ret void <i>; Return from a void function</i>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001831 ret i32 4, i8 2 <i>; Return two values 4 and 2 </i>
Chris Lattner00950542001-06-06 20:29:01 +00001832</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001833</div>
Chris Lattner00950542001-06-06 20:29:01 +00001834<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00001835<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001836<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001837<h5>Syntax:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +00001838<pre> br i1 &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 +00001839</pre>
Chris Lattner00950542001-06-06 20:29:01 +00001840<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001841<p>The '<tt>br</tt>' instruction is used to cause control flow to
1842transfer to a different basic block in the current function. There are
1843two forms of this instruction, corresponding to a conditional branch
1844and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001845<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00001846<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
Reid Spencerc78f3372007-01-12 03:35:51 +00001847single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
Reid Spencerde151942007-02-19 23:54:10 +00001848unconditional form of the '<tt>br</tt>' instruction takes a single
1849'<tt>label</tt>' value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001850<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +00001851<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00001852argument is evaluated. If the value is <tt>true</tt>, control flows
1853to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1854control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001855<h5>Example:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +00001856<pre>Test:<br> %cond = <a href="#i_icmp">icmp</a> eq, i32 %a, %b<br> br i1 %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
Reid Spencerca86e162006-12-31 07:07:53 +00001857 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001858</div>
Chris Lattner00950542001-06-06 20:29:01 +00001859<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001860<div class="doc_subsubsection">
1861 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1862</div>
1863
Misha Brukman9d0919f2003-11-08 01:05:38 +00001864<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00001865<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001866
1867<pre>
1868 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1869</pre>
1870
Chris Lattner00950542001-06-06 20:29:01 +00001871<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001872
1873<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1874several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +00001875instruction, allowing a branch to occur to one of many possible
1876destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001877
1878
Chris Lattner00950542001-06-06 20:29:01 +00001879<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001880
1881<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1882comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1883an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1884table is not allowed to contain duplicate constant entries.</p>
1885
Chris Lattner00950542001-06-06 20:29:01 +00001886<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001887
Chris Lattner261efe92003-11-25 01:02:51 +00001888<p>The <tt>switch</tt> instruction specifies a table of values and
1889destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +00001890table is searched for the given value. If the value is found, control flow is
1891transfered to the corresponding destination; otherwise, control flow is
1892transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001893
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001894<h5>Implementation:</h5>
1895
1896<p>Depending on properties of the target machine and the particular
1897<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +00001898ways. For example, it could be generated as a series of chained conditional
1899branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001900
1901<h5>Example:</h5>
1902
1903<pre>
1904 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00001905 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Reid Spencerca86e162006-12-31 07:07:53 +00001906 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001907
1908 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +00001909 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +00001910
1911 <i>; Implement a jump table:</i>
Reid Spencerca86e162006-12-31 07:07:53 +00001912 switch i32 %val, label %otherwise [ i32 0, label %onzero
1913 i32 1, label %onone
1914 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00001915</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001916</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001917
Chris Lattner00950542001-06-06 20:29:01 +00001918<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001919<div class="doc_subsubsection">
1920 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1921</div>
1922
Misha Brukman9d0919f2003-11-08 01:05:38 +00001923<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001924
Chris Lattner00950542001-06-06 20:29:01 +00001925<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001926
1927<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +00001928 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; &lt;function ptr val&gt;(&lt;function args&gt;)
Chris Lattner76b8a332006-05-14 18:23:06 +00001929 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001930</pre>
1931
Chris Lattner6536cfe2002-05-06 22:08:29 +00001932<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001933
1934<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1935function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +00001936'<tt>normal</tt>' label or the
1937'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001938"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1939"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +00001940href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Devang Patelc3fc6df2008-03-10 20:49:15 +00001941continued at the dynamically nearest "exception" label. If the callee function
Devang Patel0dbb4a12008-03-11 05:51:59 +00001942returns multiple values then individual return values are only accessible through
1943a '<tt><a href="#i_getresult">getresult</a></tt>' instruction.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001944
Chris Lattner00950542001-06-06 20:29:01 +00001945<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001946
Misha Brukman9d0919f2003-11-08 01:05:38 +00001947<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001948
Chris Lattner00950542001-06-06 20:29:01 +00001949<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001950 <li>
Duncan Sands8036ca42007-03-30 12:22:09 +00001951 The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001952 convention</a> the call should use. If none is specified, the call defaults
1953 to using C calling conventions.
1954 </li>
1955 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1956 function value being invoked. In most cases, this is a direct function
1957 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1958 an arbitrary pointer to function value.
1959 </li>
1960
1961 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1962 function to be invoked. </li>
1963
1964 <li>'<tt>function args</tt>': argument list whose types match the function
1965 signature argument types. If the function signature indicates the function
1966 accepts a variable number of arguments, the extra arguments can be
1967 specified. </li>
1968
1969 <li>'<tt>normal label</tt>': the label reached when the called function
1970 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1971
1972 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1973 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1974
Chris Lattner00950542001-06-06 20:29:01 +00001975</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001976
Chris Lattner00950542001-06-06 20:29:01 +00001977<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001978
Misha Brukman9d0919f2003-11-08 01:05:38 +00001979<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001980href="#i_call">call</a></tt>' instruction in most regards. The primary
1981difference is that it establishes an association with a label, which is used by
1982the runtime library to unwind the stack.</p>
1983
1984<p>This instruction is used in languages with destructors to ensure that proper
1985cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1986exception. Additionally, this is important for implementation of
1987'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1988
Chris Lattner00950542001-06-06 20:29:01 +00001989<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00001990<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +00001991 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001992 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +00001993 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001994 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00001995</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001996</div>
Chris Lattner35eca582004-10-16 18:04:13 +00001997
1998
Chris Lattner27f71f22003-09-03 00:41:47 +00001999<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00002000
Chris Lattner261efe92003-11-25 01:02:51 +00002001<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2002Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +00002003
Misha Brukman9d0919f2003-11-08 01:05:38 +00002004<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +00002005
Chris Lattner27f71f22003-09-03 00:41:47 +00002006<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00002007<pre>
2008 unwind
2009</pre>
2010
Chris Lattner27f71f22003-09-03 00:41:47 +00002011<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00002012
2013<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2014at the first callee in the dynamic call stack which used an <a
2015href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2016primarily used to implement exception handling.</p>
2017
Chris Lattner27f71f22003-09-03 00:41:47 +00002018<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00002019
2020<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
2021immediately halt. The dynamic call stack is then searched for the first <a
2022href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2023execution continues at the "exceptional" destination block specified by the
2024<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2025dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002026</div>
Chris Lattner35eca582004-10-16 18:04:13 +00002027
2028<!-- _______________________________________________________________________ -->
2029
2030<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2031Instruction</a> </div>
2032
2033<div class="doc_text">
2034
2035<h5>Syntax:</h5>
2036<pre>
2037 unreachable
2038</pre>
2039
2040<h5>Overview:</h5>
2041
2042<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2043instruction is used to inform the optimizer that a particular portion of the
2044code is not reachable. This can be used to indicate that the code after a
2045no-return function cannot be reached, and other facts.</p>
2046
2047<h5>Semantics:</h5>
2048
2049<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2050</div>
2051
2052
2053
Chris Lattner00950542001-06-06 20:29:01 +00002054<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +00002055<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002056<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +00002057<p>Binary operators are used to do most of the computation in a
2058program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +00002059produce a single value. The operands might represent
Reid Spencer485bad12007-02-15 03:07:05 +00002060multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
Chris Lattnera58561b2004-08-12 19:12:28 +00002061The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +00002062necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002063<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002064</div>
Chris Lattner00950542001-06-06 20:29:01 +00002065<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002066<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
2067Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002068<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002069<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002070<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 +00002071</pre>
Chris Lattner00950542001-06-06 20:29:01 +00002072<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002073<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002074<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002075<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +00002076 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
Reid Spencer485bad12007-02-15 03:07:05 +00002077 This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +00002078Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002079<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002080<p>The value produced is the integer or floating point sum of the two
2081operands.</p>
Chris Lattner5ec89832008-01-28 00:36:27 +00002082<p>If an integer sum has unsigned overflow, the result returned is the
2083mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2084the result.</p>
2085<p>Because LLVM integers use a two's complement representation, this
2086instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002087<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002088<pre> &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00002089</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002090</div>
Chris Lattner00950542001-06-06 20:29:01 +00002091<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002092<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
2093Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002094<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002095<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002096<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 +00002097</pre>
Chris Lattner00950542001-06-06 20:29:01 +00002098<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002099<p>The '<tt>sub</tt>' instruction returns the difference of its two
2100operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002101<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
2102instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002103<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002104<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00002105 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00002106values.
Reid Spencer485bad12007-02-15 03:07:05 +00002107This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +00002108Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002109<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002110<p>The value produced is the integer or floating point difference of
2111the two operands.</p>
Chris Lattner5ec89832008-01-28 00:36:27 +00002112<p>If an integer difference has unsigned overflow, the result returned is the
2113mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2114the result.</p>
2115<p>Because LLVM integers use a two's complement representation, this
2116instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002117<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +00002118<pre>
2119 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00002120 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +00002121</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002122</div>
Chris Lattner00950542001-06-06 20:29:01 +00002123<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002124<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
2125Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002126<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002127<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002128<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 +00002129</pre>
Chris Lattner00950542001-06-06 20:29:01 +00002130<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002131<p>The '<tt>mul</tt>' instruction returns the product of its two
2132operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002133<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002134<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +00002135 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +00002136values.
Reid Spencer485bad12007-02-15 03:07:05 +00002137This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +00002138Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002139<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002140<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +00002141two operands.</p>
Chris Lattner5ec89832008-01-28 00:36:27 +00002142<p>If the result of an integer multiplication has unsigned overflow,
2143the result returned is the mathematical result modulo
21442<sup>n</sup>, where n is the bit width of the result.</p>
2145<p>Because LLVM integers use a two's complement representation, and the
2146result is the same width as the operands, this instruction returns the
2147correct result for both signed and unsigned integers. If a full product
2148(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2149should be sign-extended or zero-extended as appropriate to the
2150width of the full product.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002151<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002152<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00002153</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002154</div>
Chris Lattner00950542001-06-06 20:29:01 +00002155<!-- _______________________________________________________________________ -->
Reid Spencer1628cec2006-10-26 06:15:43 +00002156<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2157</a></div>
2158<div class="doc_text">
2159<h5>Syntax:</h5>
2160<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2161</pre>
2162<h5>Overview:</h5>
2163<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2164operands.</p>
2165<h5>Arguments:</h5>
2166<p>The two arguments to the '<tt>udiv</tt>' instruction must be
2167<a href="#t_integer">integer</a> values. Both arguments must have identical
Reid Spencer485bad12007-02-15 03:07:05 +00002168types. This instruction can also take <a href="#t_vector">vector</a> versions
Reid Spencer1628cec2006-10-26 06:15:43 +00002169of the values in which case the elements must be integers.</p>
2170<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +00002171<p>The value produced is the unsigned integer quotient of the two operands.</p>
2172<p>Note that unsigned integer division and signed integer division are distinct
2173operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2174<p>Division by zero leads to undefined behavior.</p>
Reid Spencer1628cec2006-10-26 06:15:43 +00002175<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002176<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00002177</pre>
2178</div>
2179<!-- _______________________________________________________________________ -->
2180<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2181</a> </div>
2182<div class="doc_text">
2183<h5>Syntax:</h5>
2184<pre> &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2185</pre>
2186<h5>Overview:</h5>
2187<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2188operands.</p>
2189<h5>Arguments:</h5>
2190<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2191<a href="#t_integer">integer</a> values. Both arguments must have identical
Reid Spencer485bad12007-02-15 03:07:05 +00002192types. This instruction can also take <a href="#t_vector">vector</a> versions
Reid Spencer1628cec2006-10-26 06:15:43 +00002193of the values in which case the elements must be integers.</p>
2194<h5>Semantics:</h5>
Chris Lattnera73afe02008-04-01 18:45:27 +00002195<p>The value produced is the signed integer quotient of the two operands rounded towards zero.</p>
Chris Lattner5ec89832008-01-28 00:36:27 +00002196<p>Note that signed integer division and unsigned integer division are distinct
2197operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2198<p>Division by zero leads to undefined behavior. Overflow also leads to
2199undefined behavior; this is a rare case, but can occur, for example,
2200by doing a 32-bit division of -2147483648 by -1.</p>
Reid Spencer1628cec2006-10-26 06:15:43 +00002201<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002202<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00002203</pre>
2204</div>
2205<!-- _______________________________________________________________________ -->
2206<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00002207Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002208<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002209<h5>Syntax:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00002210<pre> &lt;result&gt; = fdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +00002211</pre>
2212<h5>Overview:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00002213<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
Chris Lattner261efe92003-11-25 01:02:51 +00002214operands.</p>
2215<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002216<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Reid Spencer1628cec2006-10-26 06:15:43 +00002217<a href="#t_floating">floating point</a> values. Both arguments must have
Reid Spencer485bad12007-02-15 03:07:05 +00002218identical types. This instruction can also take <a href="#t_vector">vector</a>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002219versions of floating point values.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002220<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00002221<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002222<h5>Example:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00002223<pre> &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00002224</pre>
2225</div>
2226<!-- _______________________________________________________________________ -->
Reid Spencer0a783f72006-11-02 01:53:59 +00002227<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2228</div>
2229<div class="doc_text">
2230<h5>Syntax:</h5>
2231<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2232</pre>
2233<h5>Overview:</h5>
2234<p>The '<tt>urem</tt>' instruction returns the remainder from the
2235unsigned division of its two arguments.</p>
2236<h5>Arguments:</h5>
2237<p>The two arguments to the '<tt>urem</tt>' instruction must be
2238<a href="#t_integer">integer</a> values. Both arguments must have identical
Dan Gohman80176312007-11-05 23:35:22 +00002239types. This instruction can also take <a href="#t_vector">vector</a> versions
2240of the values in which case the elements must be integers.</p>
Reid Spencer0a783f72006-11-02 01:53:59 +00002241<h5>Semantics:</h5>
2242<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattnera73afe02008-04-01 18:45:27 +00002243This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner5ec89832008-01-28 00:36:27 +00002244<p>Note that unsigned integer remainder and signed integer remainder are
2245distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2246<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Reid Spencer0a783f72006-11-02 01:53:59 +00002247<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002248<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00002249</pre>
2250
2251</div>
2252<!-- _______________________________________________________________________ -->
2253<div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00002254Instruction</a> </div>
2255<div class="doc_text">
2256<h5>Syntax:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00002257<pre> &lt;result&gt; = srem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +00002258</pre>
2259<h5>Overview:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00002260<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman80176312007-11-05 23:35:22 +00002261signed division of its two operands. This instruction can also take
2262<a href="#t_vector">vector</a> versions of the values in which case
2263the elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +00002264
Chris Lattner261efe92003-11-25 01:02:51 +00002265<h5>Arguments:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00002266<p>The two arguments to the '<tt>srem</tt>' instruction must be
2267<a href="#t_integer">integer</a> values. Both arguments must have identical
2268types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002269<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00002270<p>This instruction returns the <i>remainder</i> of a division (where the result
Reid Spencerc9fdfc82007-03-24 22:23:39 +00002271has the same sign as the dividend, <tt>var1</tt>), not the <i>modulo</i>
2272operator (where the result has the same sign as the divisor, <tt>var2</tt>) of
2273a value. For more information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +00002274 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
Reid Spencerc9fdfc82007-03-24 22:23:39 +00002275Math Forum</a>. For a table of how this is implemented in various languages,
Reid Spencer64f5c6c2007-03-24 22:40:44 +00002276please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
Reid Spencerc9fdfc82007-03-24 22:23:39 +00002277Wikipedia: modulo operation</a>.</p>
Chris Lattner5ec89832008-01-28 00:36:27 +00002278<p>Note that signed integer remainder and unsigned integer remainder are
2279distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2280<p>Taking the remainder of a division by zero leads to undefined behavior.
2281Overflow also leads to undefined behavior; this is a rare case, but can occur,
2282for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2283(The remainder doesn't actually overflow, but this rule lets srem be
2284implemented using instructions that return both the result of the division
2285and the remainder.)</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002286<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002287<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00002288</pre>
2289
2290</div>
2291<!-- _______________________________________________________________________ -->
2292<div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
2293Instruction</a> </div>
2294<div class="doc_text">
2295<h5>Syntax:</h5>
2296<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2297</pre>
2298<h5>Overview:</h5>
2299<p>The '<tt>frem</tt>' instruction returns the remainder from the
2300division of its two operands.</p>
2301<h5>Arguments:</h5>
2302<p>The two arguments to the '<tt>frem</tt>' instruction must be
2303<a href="#t_floating">floating point</a> values. Both arguments must have
Dan Gohman80176312007-11-05 23:35:22 +00002304identical types. This instruction can also take <a href="#t_vector">vector</a>
2305versions of floating point values.</p>
Reid Spencer0a783f72006-11-02 01:53:59 +00002306<h5>Semantics:</h5>
Chris Lattnera73afe02008-04-01 18:45:27 +00002307<p>This instruction returns the <i>remainder</i> of a division.
2308The remainder has the same sign as the dividend.</p>
Reid Spencer0a783f72006-11-02 01:53:59 +00002309<h5>Example:</h5>
2310<pre> &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00002311</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002312</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +00002313
Reid Spencer8e11bf82007-02-02 13:57:07 +00002314<!-- ======================================================================= -->
2315<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2316Operations</a> </div>
2317<div class="doc_text">
2318<p>Bitwise binary operators are used to do various forms of
2319bit-twiddling in a program. They are generally very efficient
2320instructions and can commonly be strength reduced from other
Chris Lattnera73afe02008-04-01 18:45:27 +00002321instructions. They require two operands of the same type, execute an operation on them,
2322and produce a single value. The resulting value is the same type as its operands.</p>
Reid Spencer8e11bf82007-02-02 13:57:07 +00002323</div>
2324
Reid Spencer569f2fa2007-01-31 21:39:12 +00002325<!-- _______________________________________________________________________ -->
2326<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2327Instruction</a> </div>
2328<div class="doc_text">
2329<h5>Syntax:</h5>
2330<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2331</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002332
Reid Spencer569f2fa2007-01-31 21:39:12 +00002333<h5>Overview:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002334
Reid Spencer569f2fa2007-01-31 21:39:12 +00002335<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2336the left a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002337
Reid Spencer569f2fa2007-01-31 21:39:12 +00002338<h5>Arguments:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002339
Reid Spencer569f2fa2007-01-31 21:39:12 +00002340<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
2341 href="#t_integer">integer</a> type.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002342
Reid Spencer569f2fa2007-01-31 21:39:12 +00002343<h5>Semantics:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002344
Chris Lattnera73afe02008-04-01 18:45:27 +00002345<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup> mod 2<sup>n</sup>,
2346where n is the width of the result. If <tt>var2</tt> is (statically or dynamically) negative or
2347equal to or larger than the number of bits in <tt>var1</tt>, the result is undefined.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002348
Reid Spencer569f2fa2007-01-31 21:39:12 +00002349<h5>Example:</h5><pre>
2350 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2351 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2352 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002353 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00002354</pre>
2355</div>
2356<!-- _______________________________________________________________________ -->
2357<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2358Instruction</a> </div>
2359<div class="doc_text">
2360<h5>Syntax:</h5>
2361<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2362</pre>
2363
2364<h5>Overview:</h5>
2365<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002366operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00002367
2368<h5>Arguments:</h5>
2369<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
2370<a href="#t_integer">integer</a> type.</p>
2371
2372<h5>Semantics:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002373
Reid Spencer569f2fa2007-01-31 21:39:12 +00002374<p>This instruction always performs a logical shift right operation. The most
2375significant bits of the result will be filled with zero bits after the
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002376shift. If <tt>var2</tt> is (statically or dynamically) equal to or larger than
2377the number of bits in <tt>var1</tt>, the result is undefined.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00002378
2379<h5>Example:</h5>
2380<pre>
2381 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2382 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2383 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2384 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002385 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00002386</pre>
2387</div>
2388
Reid Spencer8e11bf82007-02-02 13:57:07 +00002389<!-- _______________________________________________________________________ -->
Reid Spencer569f2fa2007-01-31 21:39:12 +00002390<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2391Instruction</a> </div>
2392<div class="doc_text">
2393
2394<h5>Syntax:</h5>
2395<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2396</pre>
2397
2398<h5>Overview:</h5>
2399<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002400operand shifted to the right a specified number of bits with sign extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00002401
2402<h5>Arguments:</h5>
2403<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
2404<a href="#t_integer">integer</a> type.</p>
2405
2406<h5>Semantics:</h5>
2407<p>This instruction always performs an arithmetic shift right operation,
2408The most significant bits of the result will be filled with the sign bit
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002409of <tt>var1</tt>. If <tt>var2</tt> is (statically or dynamically) equal to or
2410larger than the number of bits in <tt>var1</tt>, the result is undefined.
2411</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00002412
2413<h5>Example:</h5>
2414<pre>
2415 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2416 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2417 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2418 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00002419 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00002420</pre>
2421</div>
2422
Chris Lattner00950542001-06-06 20:29:01 +00002423<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002424<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2425Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002426<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002427<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002428<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 +00002429</pre>
Chris Lattner00950542001-06-06 20:29:01 +00002430<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002431<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2432its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002433<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002434<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +00002435 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +00002436identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002437<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002438<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002439<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00002440<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +00002441<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00002442 <tbody>
2443 <tr>
2444 <td>In0</td>
2445 <td>In1</td>
2446 <td>Out</td>
2447 </tr>
2448 <tr>
2449 <td>0</td>
2450 <td>0</td>
2451 <td>0</td>
2452 </tr>
2453 <tr>
2454 <td>0</td>
2455 <td>1</td>
2456 <td>0</td>
2457 </tr>
2458 <tr>
2459 <td>1</td>
2460 <td>0</td>
2461 <td>0</td>
2462 </tr>
2463 <tr>
2464 <td>1</td>
2465 <td>1</td>
2466 <td>1</td>
2467 </tr>
2468 </tbody>
2469</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00002470</div>
Chris Lattner00950542001-06-06 20:29:01 +00002471<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002472<pre> &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
2473 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2474 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +00002475</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002476</div>
Chris Lattner00950542001-06-06 20:29:01 +00002477<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002478<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002479<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002480<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002481<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 +00002482</pre>
Chris Lattner261efe92003-11-25 01:02:51 +00002483<h5>Overview:</h5>
2484<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2485or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002486<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002487<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +00002488 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +00002489identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002490<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002491<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002492<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00002493<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00002494<table border="1" cellspacing="0" cellpadding="4">
2495 <tbody>
2496 <tr>
2497 <td>In0</td>
2498 <td>In1</td>
2499 <td>Out</td>
2500 </tr>
2501 <tr>
2502 <td>0</td>
2503 <td>0</td>
2504 <td>0</td>
2505 </tr>
2506 <tr>
2507 <td>0</td>
2508 <td>1</td>
2509 <td>1</td>
2510 </tr>
2511 <tr>
2512 <td>1</td>
2513 <td>0</td>
2514 <td>1</td>
2515 </tr>
2516 <tr>
2517 <td>1</td>
2518 <td>1</td>
2519 <td>1</td>
2520 </tr>
2521 </tbody>
2522</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00002523</div>
Chris Lattner00950542001-06-06 20:29:01 +00002524<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002525<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2526 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2527 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +00002528</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002529</div>
Chris Lattner00950542001-06-06 20:29:01 +00002530<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002531<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2532Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002533<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +00002534<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002535<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 +00002536</pre>
Chris Lattner00950542001-06-06 20:29:01 +00002537<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002538<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2539or of its two operands. The <tt>xor</tt> is used to implement the
2540"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002541<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002542<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +00002543 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +00002544identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +00002545<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002546<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002547<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00002548<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +00002549<table border="1" cellspacing="0" cellpadding="4">
2550 <tbody>
2551 <tr>
2552 <td>In0</td>
2553 <td>In1</td>
2554 <td>Out</td>
2555 </tr>
2556 <tr>
2557 <td>0</td>
2558 <td>0</td>
2559 <td>0</td>
2560 </tr>
2561 <tr>
2562 <td>0</td>
2563 <td>1</td>
2564 <td>1</td>
2565 </tr>
2566 <tr>
2567 <td>1</td>
2568 <td>0</td>
2569 <td>1</td>
2570 </tr>
2571 <tr>
2572 <td>1</td>
2573 <td>1</td>
2574 <td>0</td>
2575 </tr>
2576 </tbody>
2577</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00002578</div>
Chris Lattner261efe92003-11-25 01:02:51 +00002579<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +00002580<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002581<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2582 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2583 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2584 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00002585</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002586</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002587
Chris Lattner00950542001-06-06 20:29:01 +00002588<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002589<div class="doc_subsection">
Chris Lattner3df241e2006-04-08 23:07:04 +00002590 <a name="vectorops">Vector Operations</a>
2591</div>
2592
2593<div class="doc_text">
2594
2595<p>LLVM supports several instructions to represent vector operations in a
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002596target-independent manner. These instructions cover the element-access and
Chris Lattner3df241e2006-04-08 23:07:04 +00002597vector-specific operations needed to process vectors effectively. While LLVM
2598does directly support these vector operations, many sophisticated algorithms
2599will want to use target-specific intrinsics to take full advantage of a specific
2600target.</p>
2601
2602</div>
2603
2604<!-- _______________________________________________________________________ -->
2605<div class="doc_subsubsection">
2606 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2607</div>
2608
2609<div class="doc_text">
2610
2611<h5>Syntax:</h5>
2612
2613<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002614 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00002615</pre>
2616
2617<h5>Overview:</h5>
2618
2619<p>
2620The '<tt>extractelement</tt>' instruction extracts a single scalar
Reid Spencer485bad12007-02-15 03:07:05 +00002621element from a vector at a specified index.
Chris Lattner3df241e2006-04-08 23:07:04 +00002622</p>
2623
2624
2625<h5>Arguments:</h5>
2626
2627<p>
2628The first operand of an '<tt>extractelement</tt>' instruction is a
Reid Spencer485bad12007-02-15 03:07:05 +00002629value of <a href="#t_vector">vector</a> type. The second operand is
Chris Lattner3df241e2006-04-08 23:07:04 +00002630an index indicating the position from which to extract the element.
2631The index may be a variable.</p>
2632
2633<h5>Semantics:</h5>
2634
2635<p>
2636The result is a scalar of the same type as the element type of
2637<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2638<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2639results are undefined.
2640</p>
2641
2642<h5>Example:</h5>
2643
2644<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002645 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00002646</pre>
2647</div>
2648
2649
2650<!-- _______________________________________________________________________ -->
2651<div class="doc_subsubsection">
2652 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2653</div>
2654
2655<div class="doc_text">
2656
2657<h5>Syntax:</h5>
2658
2659<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002660 &lt;result&gt; = insertelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt, i32 &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00002661</pre>
2662
2663<h5>Overview:</h5>
2664
2665<p>
2666The '<tt>insertelement</tt>' instruction inserts a scalar
Reid Spencer485bad12007-02-15 03:07:05 +00002667element into a vector at a specified index.
Chris Lattner3df241e2006-04-08 23:07:04 +00002668</p>
2669
2670
2671<h5>Arguments:</h5>
2672
2673<p>
2674The first operand of an '<tt>insertelement</tt>' instruction is a
Reid Spencer485bad12007-02-15 03:07:05 +00002675value of <a href="#t_vector">vector</a> type. The second operand is a
Chris Lattner3df241e2006-04-08 23:07:04 +00002676scalar value whose type must equal the element type of the first
2677operand. The third operand is an index indicating the position at
2678which to insert the value. The index may be a variable.</p>
2679
2680<h5>Semantics:</h5>
2681
2682<p>
Reid Spencer485bad12007-02-15 03:07:05 +00002683The result is a vector of the same type as <tt>val</tt>. Its
Chris Lattner3df241e2006-04-08 23:07:04 +00002684element values are those of <tt>val</tt> except at position
2685<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2686exceeds the length of <tt>val</tt>, the results are undefined.
2687</p>
2688
2689<h5>Example:</h5>
2690
2691<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002692 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00002693</pre>
2694</div>
2695
2696<!-- _______________________________________________________________________ -->
2697<div class="doc_subsubsection">
2698 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2699</div>
2700
2701<div class="doc_text">
2702
2703<h5>Syntax:</h5>
2704
2705<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002706 &lt;result&gt; = shufflevector &lt;n x &lt;ty&gt;&gt; &lt;v1&gt;, &lt;n x &lt;ty&gt;&gt; &lt;v2&gt;, &lt;n x i32&gt; &lt;mask&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00002707</pre>
2708
2709<h5>Overview:</h5>
2710
2711<p>
2712The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2713from two input vectors, returning a vector of the same type.
2714</p>
2715
2716<h5>Arguments:</h5>
2717
2718<p>
2719The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2720with types that match each other and types that match the result of the
2721instruction. The third argument is a shuffle mask, which has the same number
Reid Spencerca86e162006-12-31 07:07:53 +00002722of elements as the other vector type, but whose element type is always 'i32'.
Chris Lattner3df241e2006-04-08 23:07:04 +00002723</p>
2724
2725<p>
2726The shuffle mask operand is required to be a constant vector with either
2727constant integer or undef values.
2728</p>
2729
2730<h5>Semantics:</h5>
2731
2732<p>
2733The elements of the two input vectors are numbered from left to right across
2734both of the vectors. The shuffle mask operand specifies, for each element of
2735the result vector, which element of the two input registers the result element
2736gets. The element selector may be undef (meaning "don't care") and the second
2737operand may be undef if performing a shuffle from only one vector.
2738</p>
2739
2740<h5>Example:</h5>
2741
2742<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002743 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002744 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
Reid Spencerca86e162006-12-31 07:07:53 +00002745 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2746 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i> - Identity shuffle.
Chris Lattner3df241e2006-04-08 23:07:04 +00002747</pre>
2748</div>
2749
Tanya Lattner09474292006-04-14 19:24:33 +00002750
Chris Lattner3df241e2006-04-08 23:07:04 +00002751<!-- ======================================================================= -->
2752<div class="doc_subsection">
Chris Lattner884a9702006-08-15 00:45:58 +00002753 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002754</div>
2755
Misha Brukman9d0919f2003-11-08 01:05:38 +00002756<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002757
Chris Lattner261efe92003-11-25 01:02:51 +00002758<p>A key design point of an SSA-based representation is how it
2759represents memory. In LLVM, no memory locations are in SSA form, which
2760makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +00002761allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002762
Misha Brukman9d0919f2003-11-08 01:05:38 +00002763</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002764
Chris Lattner00950542001-06-06 20:29:01 +00002765<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002766<div class="doc_subsubsection">
2767 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2768</div>
2769
Misha Brukman9d0919f2003-11-08 01:05:38 +00002770<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002771
Chris Lattner00950542001-06-06 20:29:01 +00002772<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002773
2774<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002775 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00002776</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002777
Chris Lattner00950542001-06-06 20:29:01 +00002778<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002779
Chris Lattner261efe92003-11-25 01:02:51 +00002780<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lamb303dae92007-12-17 01:00:21 +00002781heap and returns a pointer to it. The object is always allocated in the generic
2782address space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002783
Chris Lattner00950542001-06-06 20:29:01 +00002784<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002785
2786<p>The '<tt>malloc</tt>' instruction allocates
2787<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +00002788bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +00002789appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif1acd2ee2008-02-09 22:24:34 +00002790number of elements allocated, otherwise "NumElements" is defaulted to be one.
2791If an alignment is specified, the value result of the allocation is guaranteed to
2792be aligned to at least that boundary. If not specified, or if zero, the target can
2793choose to align the allocation on any convenient boundary.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002794
Misha Brukman9d0919f2003-11-08 01:05:38 +00002795<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002796
Chris Lattner00950542001-06-06 20:29:01 +00002797<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002798
Chris Lattner261efe92003-11-25 01:02:51 +00002799<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2800a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002801
Chris Lattner2cbdc452005-11-06 08:02:57 +00002802<h5>Example:</h5>
2803
2804<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002805 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002806
Bill Wendlingaac388b2007-05-29 09:42:13 +00002807 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
2808 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
2809 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
2810 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
2811 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +00002812</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002813</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002814
Chris Lattner00950542001-06-06 20:29:01 +00002815<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002816<div class="doc_subsubsection">
2817 <a name="i_free">'<tt>free</tt>' Instruction</a>
2818</div>
2819
Misha Brukman9d0919f2003-11-08 01:05:38 +00002820<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002821
Chris Lattner00950542001-06-06 20:29:01 +00002822<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002823
2824<pre>
2825 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +00002826</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002827
Chris Lattner00950542001-06-06 20:29:01 +00002828<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002829
Chris Lattner261efe92003-11-25 01:02:51 +00002830<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +00002831memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002832
Chris Lattner00950542001-06-06 20:29:01 +00002833<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002834
Chris Lattner261efe92003-11-25 01:02:51 +00002835<p>'<tt>value</tt>' shall be a pointer value that points to a value
2836that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2837instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002838
Chris Lattner00950542001-06-06 20:29:01 +00002839<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002840
John Criswell9e2485c2004-12-10 15:51:16 +00002841<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +00002842after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002843
Chris Lattner00950542001-06-06 20:29:01 +00002844<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002845
2846<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002847 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
2848 free [4 x i8]* %array
Chris Lattner00950542001-06-06 20:29:01 +00002849</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002850</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002851
Chris Lattner00950542001-06-06 20:29:01 +00002852<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +00002853<div class="doc_subsubsection">
2854 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2855</div>
2856
Misha Brukman9d0919f2003-11-08 01:05:38 +00002857<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +00002858
Chris Lattner00950542001-06-06 20:29:01 +00002859<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002860
2861<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002862 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00002863</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002864
Chris Lattner00950542001-06-06 20:29:01 +00002865<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002866
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002867<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
2868currently executing function, to be automatically released when this function
Christopher Lamb303dae92007-12-17 01:00:21 +00002869returns to its caller. The object is always allocated in the generic address
2870space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002871
Chris Lattner00950542001-06-06 20:29:01 +00002872<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002873
John Criswell9e2485c2004-12-10 15:51:16 +00002874<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +00002875bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif1acd2ee2008-02-09 22:24:34 +00002876appropriate type to the program. If "NumElements" is specified, it is the
2877number of elements allocated, otherwise "NumElements" is defaulted to be one.
2878If an alignment is specified, the value result of the allocation is guaranteed
2879to be aligned to at least that boundary. If not specified, or if zero, the target
2880can choose to align the allocation on any convenient boundary.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002881
Misha Brukman9d0919f2003-11-08 01:05:38 +00002882<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002883
Chris Lattner00950542001-06-06 20:29:01 +00002884<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002885
John Criswellc1f786c2005-05-13 22:25:59 +00002886<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +00002887memory is automatically released when the function returns. The '<tt>alloca</tt>'
2888instruction is commonly used to represent automatic variables that must
2889have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +00002890 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002891instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002892
Chris Lattner00950542001-06-06 20:29:01 +00002893<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002894
2895<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00002896 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00002897 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
2898 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
Reid Spencerca86e162006-12-31 07:07:53 +00002899 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00002900</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002901</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00002902
Chris Lattner00950542001-06-06 20:29:01 +00002903<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002904<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2905Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002906<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00002907<h5>Syntax:</h5>
Christopher Lamb2330e4d2007-04-21 08:16:25 +00002908<pre> &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]<br> &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]<br></pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002909<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002910<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002911<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002912<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +00002913address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +00002914 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +00002915marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +00002916the number or order of execution of this <tt>load</tt> with other
2917volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2918instructions. </p>
Chris Lattnera31d1d72008-01-06 21:04:43 +00002919<p>
2920The optional "align" argument specifies the alignment of the operation
2921(that is, the alignment of the memory address). A value of 0 or an
2922omitted "align" argument means that the operation has the preferential
2923alignment for the target. It is the responsibility of the code emitter
2924to ensure that the alignment information is correct. Overestimating
2925the alignment results in an undefined behavior. Underestimating the
2926alignment may produce less efficient code. An alignment of 1 is always
2927safe.
2928</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002929<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002930<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002931<h5>Examples:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002932<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Chris Lattner261efe92003-11-25 01:02:51 +00002933 <a
Reid Spencerca86e162006-12-31 07:07:53 +00002934 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2935 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002936</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002937</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002938<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +00002939<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2940Instruction</a> </div>
Reid Spencer035ab572006-11-09 21:18:01 +00002941<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +00002942<h5>Syntax:</h5>
Christopher Lamb2330e4d2007-04-21 08:16:25 +00002943<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
2944 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002945</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002946<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002947<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002948<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00002949<p>There are two arguments to the '<tt>store</tt>' instruction: a value
Jeff Cohen6f1cc772007-04-22 01:17:39 +00002950to store and an address at which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +00002951operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +00002952operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +00002953optimizer is not allowed to modify the number or order of execution of
2954this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2955 href="#i_store">store</a></tt> instructions.</p>
Chris Lattnera31d1d72008-01-06 21:04:43 +00002956<p>
2957The optional "align" argument specifies the alignment of the operation
2958(that is, the alignment of the memory address). A value of 0 or an
2959omitted "align" argument means that the operation has the preferential
2960alignment for the target. It is the responsibility of the code emitter
2961to ensure that the alignment information is correct. Overestimating
2962the alignment results in an undefined behavior. Underestimating the
2963alignment may produce less efficient code. An alignment of 1 is always
2964safe.
2965</p>
Chris Lattner261efe92003-11-25 01:02:51 +00002966<h5>Semantics:</h5>
2967<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2968at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002969<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00002970<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +00002971 store i32 3, i32* %ptr <i>; yields {void}</i>
2972 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00002973</pre>
Reid Spencer47ce1792006-11-09 21:15:49 +00002974</div>
2975
Chris Lattner2b7d3202002-05-06 03:03:22 +00002976<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002977<div class="doc_subsubsection">
2978 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2979</div>
2980
Misha Brukman9d0919f2003-11-08 01:05:38 +00002981<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +00002982<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002983<pre>
2984 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2985</pre>
2986
Chris Lattner7faa8832002-04-14 06:13:44 +00002987<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002988
2989<p>
2990The '<tt>getelementptr</tt>' instruction is used to get the address of a
2991subelement of an aggregate data structure.</p>
2992
Chris Lattner7faa8832002-04-14 06:13:44 +00002993<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002994
Reid Spencer85f5b5b2006-12-04 21:29:24 +00002995<p>This instruction takes a list of integer operands that indicate what
Chris Lattnerf74d5c72004-04-05 01:30:49 +00002996elements of the aggregate object to index to. The actual types of the arguments
2997provided depend on the type of the first pointer argument. The
2998'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +00002999levels of a structure or to a specific index in an array. When indexing into a
Reid Spencerca86e162006-12-31 07:07:53 +00003000structure, only <tt>i32</tt> integer constants are allowed. When indexing
Reid Spencer85f5b5b2006-12-04 21:29:24 +00003001into an array or pointer, only integers of 32 or 64 bits are allowed, and will
3002be sign extended to 64-bit values.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003003
Chris Lattner261efe92003-11-25 01:02:51 +00003004<p>For example, let's consider a C code fragment and how it gets
3005compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003006
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003007<div class="doc_code">
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003008<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003009struct RT {
3010 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +00003011 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003012 char C;
3013};
3014struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +00003015 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003016 double Y;
3017 struct RT Z;
3018};
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003019
Chris Lattnercabc8462007-05-29 15:43:56 +00003020int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003021 return &amp;s[1].Z.B[5][13];
3022}
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003023</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003024</div>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003025
Misha Brukman9d0919f2003-11-08 01:05:38 +00003026<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003027
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003028<div class="doc_code">
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003029<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003030%RT = type { i8 , [10 x [20 x i32]], i8 }
3031%ST = type { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003032
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003033define i32* %foo(%ST* %s) {
3034entry:
3035 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3036 ret i32* %reg
3037}
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003038</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00003039</div>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003040
Chris Lattner7faa8832002-04-14 06:13:44 +00003041<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003042
3043<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +00003044on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Reid Spencer85f5b5b2006-12-04 21:29:24 +00003045and <a href="#t_array">array</a> types can use a 32-bit or 64-bit
Reid Spencer42ddd842006-12-03 16:53:48 +00003046<a href="#t_integer">integer</a> type but the value will always be sign extended
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003047to 64-bits. <a href="#t_struct">Structure</a> types require <tt>i32</tt>
Reid Spencer42ddd842006-12-03 16:53:48 +00003048<b>constants</b>.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003049
Misha Brukman9d0919f2003-11-08 01:05:38 +00003050<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Reid Spencerca86e162006-12-31 07:07:53 +00003051type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003052}</tt>' type, a structure. The second index indexes into the third element of
Reid Spencerca86e162006-12-31 07:07:53 +00003053the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3054i8 }</tt>' type, another structure. The third index indexes into the second
3055element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003056array. The two dimensions of the array are subscripted into, yielding an
Reid Spencerca86e162006-12-31 07:07:53 +00003057'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3058to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003059
Chris Lattner261efe92003-11-25 01:02:51 +00003060<p>Note that it is perfectly legal to index partially through a
3061structure, returning a pointer to an inner element. Because of this,
3062the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003063
3064<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00003065 define i32* %foo(%ST* %s) {
3066 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003067 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3068 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003069 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3070 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3071 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003072 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00003073</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00003074
3075<p>Note that it is undefined to access an array out of bounds: array and
3076pointer indexes must always be within the defined bounds of the array type.
3077The one exception for this rules is zero length arrays. These arrays are
3078defined to be accessible as variable length arrays, which requires access
3079beyond the zero'th element.</p>
3080
Chris Lattner884a9702006-08-15 00:45:58 +00003081<p>The getelementptr instruction is often confusing. For some more insight
3082into how it works, see <a href="GetElementPtr.html">the getelementptr
3083FAQ</a>.</p>
3084
Chris Lattner7faa8832002-04-14 06:13:44 +00003085<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +00003086
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003087<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00003088 <i>; yields [12 x i8]*:aptr</i>
3089 %aptr = getelementptr {i32, [12 x i8]}* %sptr, i64 0, i32 1
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003090</pre>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00003091</div>
Reid Spencer47ce1792006-11-09 21:15:49 +00003092
Chris Lattner00950542001-06-06 20:29:01 +00003093<!-- ======================================================================= -->
Reid Spencer2fd21e62006-11-08 01:18:52 +00003094<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003095</div>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003096<div class="doc_text">
Reid Spencer2fd21e62006-11-08 01:18:52 +00003097<p>The instructions in this category are the conversion instructions (casting)
3098which all take a single operand and a type. They perform various bit conversions
3099on the operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003100</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003101
Chris Lattner6536cfe2002-05-06 22:08:29 +00003102<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +00003103<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003104 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3105</div>
3106<div class="doc_text">
3107
3108<h5>Syntax:</h5>
3109<pre>
3110 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3111</pre>
3112
3113<h5>Overview:</h5>
3114<p>
3115The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3116</p>
3117
3118<h5>Arguments:</h5>
3119<p>
3120The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3121be an <a href="#t_integer">integer</a> type, and a type that specifies the size
Chris Lattner3b19d652007-01-15 01:54:13 +00003122and type of the result, which must be an <a href="#t_integer">integer</a>
Reid Spencerd4448792006-11-09 23:03:26 +00003123type. The bit size of <tt>value</tt> must be larger than the bit size of
3124<tt>ty2</tt>. Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003125
3126<h5>Semantics:</h5>
3127<p>
3128The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
Reid Spencerd4448792006-11-09 23:03:26 +00003129and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3130larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3131It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003132
3133<h5>Example:</h5>
3134<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00003135 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00003136 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3137 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003138</pre>
3139</div>
3140
3141<!-- _______________________________________________________________________ -->
3142<div class="doc_subsubsection">
3143 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3144</div>
3145<div class="doc_text">
3146
3147<h5>Syntax:</h5>
3148<pre>
3149 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3150</pre>
3151
3152<h5>Overview:</h5>
3153<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3154<tt>ty2</tt>.</p>
3155
3156
3157<h5>Arguments:</h5>
3158<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Chris Lattner3b19d652007-01-15 01:54:13 +00003159<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3160also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencerd4448792006-11-09 23:03:26 +00003161<tt>value</tt> must be smaller than the bit size of the destination type,
3162<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003163
3164<h5>Semantics:</h5>
3165<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Chris Lattnerd1d25172007-05-24 19:13:27 +00003166bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003167
Reid Spencerb5929522007-01-12 15:46:11 +00003168<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003169
3170<h5>Example:</h5>
3171<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00003172 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00003173 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003174</pre>
3175</div>
3176
3177<!-- _______________________________________________________________________ -->
3178<div class="doc_subsubsection">
3179 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3180</div>
3181<div class="doc_text">
3182
3183<h5>Syntax:</h5>
3184<pre>
3185 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3186</pre>
3187
3188<h5>Overview:</h5>
3189<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3190
3191<h5>Arguments:</h5>
3192<p>
3193The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Chris Lattner3b19d652007-01-15 01:54:13 +00003194<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3195also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencerd4448792006-11-09 23:03:26 +00003196<tt>value</tt> must be smaller than the bit size of the destination type,
3197<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003198
3199<h5>Semantics:</h5>
3200<p>
3201The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3202bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
Chris Lattnerd1d25172007-05-24 19:13:27 +00003203the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003204
Reid Spencerc78f3372007-01-12 03:35:51 +00003205<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003206
3207<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003208<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00003209 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00003210 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003211</pre>
3212</div>
3213
3214<!-- _______________________________________________________________________ -->
3215<div class="doc_subsubsection">
Reid Spencer3fa91b02006-11-09 21:48:10 +00003216 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3217</div>
3218
3219<div class="doc_text">
3220
3221<h5>Syntax:</h5>
3222
3223<pre>
3224 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3225</pre>
3226
3227<h5>Overview:</h5>
3228<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3229<tt>ty2</tt>.</p>
3230
3231
3232<h5>Arguments:</h5>
3233<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3234 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3235cast it to. The size of <tt>value</tt> must be larger than the size of
3236<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3237<i>no-op cast</i>.</p>
3238
3239<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003240<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3241<a href="#t_floating">floating point</a> type to a smaller
3242<a href="#t_floating">floating point</a> type. If the value cannot fit within
3243the destination type, <tt>ty2</tt>, then the results are undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00003244
3245<h5>Example:</h5>
3246<pre>
3247 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3248 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3249</pre>
3250</div>
3251
3252<!-- _______________________________________________________________________ -->
3253<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003254 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3255</div>
3256<div class="doc_text">
3257
3258<h5>Syntax:</h5>
3259<pre>
3260 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3261</pre>
3262
3263<h5>Overview:</h5>
3264<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3265floating point value.</p>
3266
3267<h5>Arguments:</h5>
3268<p>The '<tt>fpext</tt>' instruction takes a
3269<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
Reid Spencerd4448792006-11-09 23:03:26 +00003270and a <a href="#t_floating">floating point</a> type to cast it to. The source
3271type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003272
3273<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003274<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Duncan Sands8036ca42007-03-30 12:22:09 +00003275<a href="#t_floating">floating point</a> type to a larger
3276<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
Reid Spencerd4448792006-11-09 23:03:26 +00003277used to make a <i>no-op cast</i> because it always changes bits. Use
Reid Spencer5c0ef472006-11-11 23:08:07 +00003278<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003279
3280<h5>Example:</h5>
3281<pre>
3282 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3283 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3284</pre>
3285</div>
3286
3287<!-- _______________________________________________________________________ -->
3288<div class="doc_subsubsection">
Reid Spencer24d6da52007-01-21 00:29:26 +00003289 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003290</div>
3291<div class="doc_text">
3292
3293<h5>Syntax:</h5>
3294<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +00003295 &lt;result&gt; = fptoui &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003296</pre>
3297
3298<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +00003299<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003300unsigned integer equivalent of type <tt>ty2</tt>.
3301</p>
3302
3303<h5>Arguments:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +00003304<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
Nate Begemanb348d182007-11-17 03:58:34 +00003305scalar or vector <a href="#t_floating">floating point</a> value, and a type
3306to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3307type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3308vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003309
3310<h5>Semantics:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +00003311<p> The '<tt>fptoui</tt>' instruction converts its
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003312<a href="#t_floating">floating point</a> operand into the nearest (rounding
3313towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3314the results are undefined.</p>
3315
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003316<h5>Example:</h5>
3317<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +00003318 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +00003319 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencer1539a1c2007-07-31 14:40:14 +00003320 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003321</pre>
3322</div>
3323
3324<!-- _______________________________________________________________________ -->
3325<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +00003326 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003327</div>
3328<div class="doc_text">
3329
3330<h5>Syntax:</h5>
3331<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00003332 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003333</pre>
3334
3335<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003336<p>The '<tt>fptosi</tt>' instruction converts
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003337<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
Chris Lattnercc37aae2004-03-12 05:50:16 +00003338</p>
3339
Chris Lattner6536cfe2002-05-06 22:08:29 +00003340<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003341<p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
Nate Begemanb348d182007-11-17 03:58:34 +00003342scalar or vector <a href="#t_floating">floating point</a> value, and a type
3343to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3344type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3345vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003346
Chris Lattner6536cfe2002-05-06 22:08:29 +00003347<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003348<p>The '<tt>fptosi</tt>' instruction converts its
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003349<a href="#t_floating">floating point</a> operand into the nearest (rounding
3350towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3351the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003352
Chris Lattner33ba0d92001-07-09 00:26:23 +00003353<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003354<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +00003355 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +00003356 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003357 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003358</pre>
3359</div>
3360
3361<!-- _______________________________________________________________________ -->
3362<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +00003363 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003364</div>
3365<div class="doc_text">
3366
3367<h5>Syntax:</h5>
3368<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00003369 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003370</pre>
3371
3372<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003373<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003374integer and converts that value to the <tt>ty2</tt> type.</p>
3375
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003376<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +00003377<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3378scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3379to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3380type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3381floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003382
3383<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003384<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003385integer quantity and converts it to the corresponding floating point value. If
Jeff Cohencb757312007-04-22 14:56:37 +00003386the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003387
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003388<h5>Example:</h5>
3389<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00003390 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003391 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003392</pre>
3393</div>
3394
3395<!-- _______________________________________________________________________ -->
3396<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +00003397 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003398</div>
3399<div class="doc_text">
3400
3401<h5>Syntax:</h5>
3402<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00003403 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003404</pre>
3405
3406<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003407<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003408integer and converts that value to the <tt>ty2</tt> type.</p>
3409
3410<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +00003411<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3412scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3413to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3414type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3415floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003416
3417<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00003418<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003419integer quantity and converts it to the corresponding floating point value. If
Jeff Cohencb757312007-04-22 14:56:37 +00003420the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003421
3422<h5>Example:</h5>
3423<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00003424 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003425 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003426</pre>
3427</div>
3428
3429<!-- _______________________________________________________________________ -->
3430<div class="doc_subsubsection">
Reid Spencer72679252006-11-11 21:00:47 +00003431 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3432</div>
3433<div class="doc_text">
3434
3435<h5>Syntax:</h5>
3436<pre>
3437 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3438</pre>
3439
3440<h5>Overview:</h5>
3441<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3442the integer type <tt>ty2</tt>.</p>
3443
3444<h5>Arguments:</h5>
3445<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Duncan Sands8036ca42007-03-30 12:22:09 +00003446must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
Reid Spencer72679252006-11-11 21:00:47 +00003447<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3448
3449<h5>Semantics:</h5>
3450<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3451<tt>ty2</tt> by interpreting the pointer value as an integer and either
3452truncating or zero extending that value to the size of the integer type. If
3453<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3454<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
Jeff Cohenb627eab2007-04-29 01:07:00 +00003455are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3456change.</p>
Reid Spencer72679252006-11-11 21:00:47 +00003457
3458<h5>Example:</h5>
3459<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003460 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3461 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +00003462</pre>
3463</div>
3464
3465<!-- _______________________________________________________________________ -->
3466<div class="doc_subsubsection">
3467 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3468</div>
3469<div class="doc_text">
3470
3471<h5>Syntax:</h5>
3472<pre>
3473 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3474</pre>
3475
3476<h5>Overview:</h5>
3477<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3478a pointer type, <tt>ty2</tt>.</p>
3479
3480<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +00003481<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Reid Spencer72679252006-11-11 21:00:47 +00003482value to cast, and a type to cast it to, which must be a
Anton Korobeynikov7f705592007-01-12 19:20:47 +00003483<a href="#t_pointer">pointer</a> type.
Reid Spencer72679252006-11-11 21:00:47 +00003484
3485<h5>Semantics:</h5>
3486<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3487<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3488the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3489size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3490the size of a pointer then a zero extension is done. If they are the same size,
3491nothing is done (<i>no-op cast</i>).</p>
3492
3493<h5>Example:</h5>
3494<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003495 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3496 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3497 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +00003498</pre>
3499</div>
3500
3501<!-- _______________________________________________________________________ -->
3502<div class="doc_subsubsection">
Reid Spencer5c0ef472006-11-11 23:08:07 +00003503 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003504</div>
3505<div class="doc_text">
3506
3507<h5>Syntax:</h5>
3508<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003509 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003510</pre>
3511
3512<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003513<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003514<tt>ty2</tt> without changing any bits.</p>
3515
3516<h5>Arguments:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003517<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003518a first class value, and a type to cast it to, which must also be a <a
3519 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
Reid Spencer19b569f2007-01-09 20:08:58 +00003520and the destination type, <tt>ty2</tt>, must be identical. If the source
3521type is a pointer, the destination type must also be a pointer.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003522
3523<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00003524<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer72679252006-11-11 21:00:47 +00003525<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3526this conversion. The conversion is done as if the <tt>value</tt> had been
3527stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3528converted to other pointer types with this instruction. To convert pointers to
3529other types, use the <a href="#i_inttoptr">inttoptr</a> or
3530<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00003531
3532<h5>Example:</h5>
3533<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003534 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003535 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3536 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00003537</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003538</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003539
Reid Spencer2fd21e62006-11-08 01:18:52 +00003540<!-- ======================================================================= -->
3541<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3542<div class="doc_text">
3543<p>The instructions in this category are the "miscellaneous"
3544instructions, which defy better classification.</p>
3545</div>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003546
3547<!-- _______________________________________________________________________ -->
3548<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3549</div>
3550<div class="doc_text">
3551<h5>Syntax:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003552<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003553</pre>
3554<h5>Overview:</h5>
3555<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
3556of its two integer operands.</p>
3557<h5>Arguments:</h5>
3558<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Jeff Cohenb627eab2007-04-29 01:07:00 +00003559the condition code indicating the kind of comparison to perform. It is not
3560a value, just a keyword. The possible condition code are:
Reid Spencerf3a70a62006-11-18 21:50:54 +00003561<ol>
3562 <li><tt>eq</tt>: equal</li>
3563 <li><tt>ne</tt>: not equal </li>
3564 <li><tt>ugt</tt>: unsigned greater than</li>
3565 <li><tt>uge</tt>: unsigned greater or equal</li>
3566 <li><tt>ult</tt>: unsigned less than</li>
3567 <li><tt>ule</tt>: unsigned less or equal</li>
3568 <li><tt>sgt</tt>: signed greater than</li>
3569 <li><tt>sge</tt>: signed greater or equal</li>
3570 <li><tt>slt</tt>: signed less than</li>
3571 <li><tt>sle</tt>: signed less or equal</li>
3572</ol>
Chris Lattner3b19d652007-01-15 01:54:13 +00003573<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Reid Spencer350f8aa2007-01-04 05:19:58 +00003574<a href="#t_pointer">pointer</a> typed. They must also be identical types.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003575<h5>Semantics:</h5>
3576<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3577the condition code given as <tt>cond</tt>. The comparison performed always
Reid Spencerc78f3372007-01-12 03:35:51 +00003578yields a <a href="#t_primitive">i1</a> result, as follows:
Reid Spencerf3a70a62006-11-18 21:50:54 +00003579<ol>
3580 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3581 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3582 </li>
3583 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3584 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3585 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3586 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3587 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3588 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3589 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3590 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3591 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3592 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3593 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3594 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3595 <li><tt>sge</tt>: interprets the operands as signed values and yields
3596 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3597 <li><tt>slt</tt>: interprets the operands as signed values and yields
3598 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3599 <li><tt>sle</tt>: interprets the operands as signed values and yields
3600 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003601</ol>
3602<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Jeff Cohenb627eab2007-04-29 01:07:00 +00003603values are compared as if they were integers.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003604
3605<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00003606<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3607 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3608 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3609 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3610 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3611 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003612</pre>
3613</div>
3614
3615<!-- _______________________________________________________________________ -->
3616<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3617</div>
3618<div class="doc_text">
3619<h5>Syntax:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003620<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003621</pre>
3622<h5>Overview:</h5>
3623<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3624of its floating point operands.</p>
3625<h5>Arguments:</h5>
3626<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Jeff Cohenb627eab2007-04-29 01:07:00 +00003627the condition code indicating the kind of comparison to perform. It is not
3628a value, just a keyword. The possible condition code are:
Reid Spencerf3a70a62006-11-18 21:50:54 +00003629<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +00003630 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003631 <li><tt>oeq</tt>: ordered and equal</li>
3632 <li><tt>ogt</tt>: ordered and greater than </li>
3633 <li><tt>oge</tt>: ordered and greater than or equal</li>
3634 <li><tt>olt</tt>: ordered and less than </li>
3635 <li><tt>ole</tt>: ordered and less than or equal</li>
3636 <li><tt>one</tt>: ordered and not equal</li>
3637 <li><tt>ord</tt>: ordered (no nans)</li>
3638 <li><tt>ueq</tt>: unordered or equal</li>
3639 <li><tt>ugt</tt>: unordered or greater than </li>
3640 <li><tt>uge</tt>: unordered or greater than or equal</li>
3641 <li><tt>ult</tt>: unordered or less than </li>
3642 <li><tt>ule</tt>: unordered or less than or equal</li>
3643 <li><tt>une</tt>: unordered or not equal</li>
3644 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003645 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003646</ol>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003647<p><i>Ordered</i> means that neither operand is a QNAN while
Reid Spencer93a49852006-12-06 07:08:07 +00003648<i>unordered</i> means that either operand may be a QNAN.</p>
Reid Spencer350f8aa2007-01-04 05:19:58 +00003649<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be
3650<a href="#t_floating">floating point</a> typed. They must have identical
3651types.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003652<h5>Semantics:</h5>
3653<p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3654the condition code given as <tt>cond</tt>. The comparison performed always
Reid Spencerc78f3372007-01-12 03:35:51 +00003655yields a <a href="#t_primitive">i1</a> result, as follows:
Reid Spencerf3a70a62006-11-18 21:50:54 +00003656<ol>
3657 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003658 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003659 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003660 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003661 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003662 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003663 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003664 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003665 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003666 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003667 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003668 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +00003669 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003670 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3671 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003672 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003673 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003674 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003675 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003676 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003677 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003678 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003679 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003680 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003681 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +00003682 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00003683 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003684 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3685</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +00003686
3687<h5>Example:</h5>
3688<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3689 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3690 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3691 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3692</pre>
3693</div>
3694
Reid Spencer2fd21e62006-11-08 01:18:52 +00003695<!-- _______________________________________________________________________ -->
3696<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3697Instruction</a> </div>
3698<div class="doc_text">
3699<h5>Syntax:</h5>
3700<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
3701<h5>Overview:</h5>
3702<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
3703the SSA graph representing the function.</p>
3704<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003705<p>The type of the incoming values is specified with the first type
Reid Spencer2fd21e62006-11-08 01:18:52 +00003706field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3707as arguments, with one pair for each predecessor basic block of the
3708current block. Only values of <a href="#t_firstclass">first class</a>
3709type may be used as the value arguments to the PHI node. Only labels
3710may be used as the label arguments.</p>
3711<p>There must be no non-phi instructions between the start of a basic
3712block and the PHI instructions: i.e. PHI instructions must be first in
3713a basic block.</p>
3714<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00003715<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
3716specified by the pair corresponding to the predecessor basic block that executed
3717just prior to the current block.</p>
Reid Spencer2fd21e62006-11-08 01:18:52 +00003718<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +00003719<pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add i32 %indvar, 1<br> br label %Loop<br></pre>
Reid Spencer2fd21e62006-11-08 01:18:52 +00003720</div>
3721
Chris Lattnercc37aae2004-03-12 05:50:16 +00003722<!-- _______________________________________________________________________ -->
3723<div class="doc_subsubsection">
3724 <a name="i_select">'<tt>select</tt>' Instruction</a>
3725</div>
3726
3727<div class="doc_text">
3728
3729<h5>Syntax:</h5>
3730
3731<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +00003732 &lt;result&gt; = select i1 &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003733</pre>
3734
3735<h5>Overview:</h5>
3736
3737<p>
3738The '<tt>select</tt>' instruction is used to choose one value based on a
3739condition, without branching.
3740</p>
3741
3742
3743<h5>Arguments:</h5>
3744
3745<p>
3746The '<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.
3747</p>
3748
3749<h5>Semantics:</h5>
3750
3751<p>
3752If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +00003753value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +00003754</p>
3755
3756<h5>Example:</h5>
3757
3758<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +00003759 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +00003760</pre>
3761</div>
3762
Robert Bocchino05ccd702006-01-15 20:48:27 +00003763
3764<!-- _______________________________________________________________________ -->
3765<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +00003766 <a name="i_call">'<tt>call</tt>' Instruction</a>
3767</div>
3768
Misha Brukman9d0919f2003-11-08 01:05:38 +00003769<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +00003770
Chris Lattner00950542001-06-06 20:29:01 +00003771<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003772<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +00003773 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] &lt;ty&gt; [&lt;fnty&gt;*] &lt;fnptrval&gt;(&lt;param list&gt;)
Chris Lattner2bff5242005-05-06 05:47:36 +00003774</pre>
3775
Chris Lattner00950542001-06-06 20:29:01 +00003776<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003777
Misha Brukman9d0919f2003-11-08 01:05:38 +00003778<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003779
Chris Lattner00950542001-06-06 20:29:01 +00003780<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003781
Misha Brukman9d0919f2003-11-08 01:05:38 +00003782<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003783
Chris Lattner6536cfe2002-05-06 22:08:29 +00003784<ol>
Chris Lattner261efe92003-11-25 01:02:51 +00003785 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003786 <p>The optional "tail" marker indicates whether the callee function accesses
3787 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +00003788 function call is eligible for tail call optimization. Note that calls may
3789 be marked "tail" even if they do not occur before a <a
3790 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +00003791 </li>
3792 <li>
Duncan Sands8036ca42007-03-30 12:22:09 +00003793 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003794 convention</a> the call should use. If none is specified, the call defaults
3795 to using C calling conventions.
3796 </li>
3797 <li>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +00003798 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
3799 the type of the return value. Functions that return no value are marked
3800 <tt><a href="#t_void">void</a></tt>.</p>
3801 </li>
3802 <li>
3803 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
3804 value being invoked. The argument types must match the types implied by
3805 this signature. This type can be omitted if the function is not varargs
3806 and if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003807 </li>
3808 <li>
3809 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3810 be invoked. In most cases, this is a direct function invocation, but
3811 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +00003812 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00003813 </li>
3814 <li>
3815 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +00003816 function signature argument types. All arguments must be of
3817 <a href="#t_firstclass">first class</a> type. If the function signature
3818 indicates the function accepts a variable number of arguments, the extra
3819 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +00003820 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00003821</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00003822
Chris Lattner00950542001-06-06 20:29:01 +00003823<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003824
Chris Lattner261efe92003-11-25 01:02:51 +00003825<p>The '<tt>call</tt>' instruction is used to cause control flow to
3826transfer to a specified function, with its incoming arguments bound to
3827the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3828instruction in the called function, control flow continues with the
3829instruction after the function call, and the return value of the
Chris Lattner772fccf2008-03-21 17:24:17 +00003830function is bound to the result argument. If the callee returns multiple
3831values then the return values of the function are only accessible through
3832the '<tt><a href="#i_getresult">getresult</a></tt>' instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00003833
Chris Lattner00950542001-06-06 20:29:01 +00003834<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00003835
3836<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +00003837 %retval = call i32 @test(i32 %argc)
Chris Lattner772fccf2008-03-21 17:24:17 +00003838 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
3839 %X = tail call i32 @foo() <i>; yields i32</i>
3840 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
3841 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +00003842
3843 %struct.A = type { i32, i8 }
Chris Lattner772fccf2008-03-21 17:24:17 +00003844 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
3845 %gr = getresult %struct.A %r, 0 <i>; yields i32</i>
3846 %gr1 = getresult %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner2bff5242005-05-06 05:47:36 +00003847</pre>
3848
Misha Brukman9d0919f2003-11-08 01:05:38 +00003849</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003850
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003851<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +00003852<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +00003853 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003854</div>
3855
Misha Brukman9d0919f2003-11-08 01:05:38 +00003856<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +00003857
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003858<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003859
3860<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003861 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00003862</pre>
3863
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003864<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003865
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003866<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +00003867the "variable argument" area of a function call. It is used to implement the
3868<tt>va_arg</tt> macro in C.</p>
3869
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003870<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003871
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003872<p>This instruction takes a <tt>va_list*</tt> value and the type of
3873the argument. It returns a value of the specified argument type and
Jeff Cohenb627eab2007-04-29 01:07:00 +00003874increments the <tt>va_list</tt> to point to the next argument. The
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003875actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003876
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003877<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003878
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003879<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3880type from the specified <tt>va_list</tt> and causes the
3881<tt>va_list</tt> to point to the next argument. For more information,
3882see the variable argument handling <a href="#int_varargs">Intrinsic
3883Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003884
3885<p>It is legal for this instruction to be called in a function which does not
3886take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003887function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003888
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00003889<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +00003890href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +00003891argument.</p>
3892
Chris Lattner8d1a81d2003-10-18 05:51:36 +00003893<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00003894
3895<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3896
Misha Brukman9d0919f2003-11-08 01:05:38 +00003897</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003898
Devang Patelc3fc6df2008-03-10 20:49:15 +00003899<!-- _______________________________________________________________________ -->
3900<div class="doc_subsubsection">
3901 <a name="i_getresult">'<tt>getresult</tt>' Instruction</a>
3902</div>
3903
3904<div class="doc_text">
3905
3906<h5>Syntax:</h5>
3907<pre>
Chris Lattner45c5e8b2008-03-21 17:20:51 +00003908 &lt;resultval&gt; = getresult &lt;type&gt; &lt;retval&gt;, &lt;index&gt;
Devang Patelc3fc6df2008-03-10 20:49:15 +00003909</pre>
Chris Lattner45c5e8b2008-03-21 17:20:51 +00003910
Devang Patelc3fc6df2008-03-10 20:49:15 +00003911<h5>Overview:</h5>
3912
3913<p> The '<tt>getresult</tt>' instruction is used to extract individual values
Chris Lattner45c5e8b2008-03-21 17:20:51 +00003914from a '<tt><a href="#i_call">call</a></tt>'
3915or '<tt><a href="#i_invoke">invoke</a></tt>' instruction that returns multiple
3916results.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +00003917
3918<h5>Arguments:</h5>
3919
Chris Lattner45c5e8b2008-03-21 17:20:51 +00003920<p>The '<tt>getresult</tt>' instruction takes a call or invoke value as its
3921first argument. The value must have <a href="#t_struct">structure type</a>.
3922The second argument is an unsigned index value which must be in range for
3923the number of values returned by the call.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +00003924
3925<h5>Semantics:</h5>
3926
Chris Lattner45c5e8b2008-03-21 17:20:51 +00003927<p>The '<tt>getresult</tt>' instruction extracts the element identified by
3928'<tt>index</tt>' from the aggregate value.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +00003929
3930<h5>Example:</h5>
3931
3932<pre>
3933 %struct.A = type { i32, i8 }
3934
3935 %r = call %struct.A @foo()
Chris Lattner45c5e8b2008-03-21 17:20:51 +00003936 %gr = getresult %struct.A %r, 0 <i>; yields i32:%gr</i>
3937 %gr1 = getresult %struct.A %r, 1 <i>; yields i8:%gr1</i>
Devang Patelc3fc6df2008-03-10 20:49:15 +00003938 add i32 %gr, 42
3939 add i8 %gr1, 41
3940</pre>
3941
3942</div>
3943
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003944<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +00003945<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3946<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00003947
Misha Brukman9d0919f2003-11-08 01:05:38 +00003948<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +00003949
3950<p>LLVM supports the notion of an "intrinsic function". These functions have
Reid Spencer409e28f2007-04-01 08:04:23 +00003951well known names and semantics and are required to follow certain restrictions.
3952Overall, these intrinsics represent an extension mechanism for the LLVM
Jeff Cohenb627eab2007-04-29 01:07:00 +00003953language that does not require changing all of the transformations in LLVM when
Gabor Greif04367bf2007-07-06 22:07:22 +00003954adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00003955
John Criswellfc6b8952005-05-16 16:17:45 +00003956<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Jeff Cohenb627eab2007-04-29 01:07:00 +00003957prefix is reserved in LLVM for intrinsic names; thus, function names may not
3958begin with this prefix. Intrinsic functions must always be external functions:
3959you cannot define the body of intrinsic functions. Intrinsic functions may
3960only be used in call or invoke instructions: it is illegal to take the address
3961of an intrinsic function. Additionally, because intrinsic functions are part
3962of the LLVM language, it is required if any are added that they be documented
3963here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00003964
Chandler Carruth69940402007-08-04 01:51:18 +00003965<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
3966a family of functions that perform the same operation but on different data
3967types. Because LLVM can represent over 8 million different integer types,
3968overloading is used commonly to allow an intrinsic function to operate on any
3969integer type. One or more of the argument types or the result type can be
3970overloaded to accept any integer type. Argument types may also be defined as
3971exactly matching a previous argument's type or the result type. This allows an
3972intrinsic function which accepts multiple arguments, but needs all of them to
3973be of the same type, to only be overloaded with respect to a single argument or
3974the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00003975
Chandler Carruth69940402007-08-04 01:51:18 +00003976<p>Overloaded intrinsics will have the names of its overloaded argument types
3977encoded into its function name, each preceded by a period. Only those types
3978which are overloaded result in a name suffix. Arguments whose type is matched
3979against another type do not. For example, the <tt>llvm.ctpop</tt> function can
3980take an integer of any width and returns an integer of exactly the same integer
3981width. This leads to a family of functions such as
3982<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
3983Only one type, the return type, is overloaded, and only one type suffix is
3984required. Because the argument's type is matched against the return type, it
3985does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +00003986
3987<p>To learn how to add an intrinsic function, please see the
3988<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +00003989</p>
3990
Misha Brukman9d0919f2003-11-08 01:05:38 +00003991</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00003992
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00003993<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +00003994<div class="doc_subsection">
3995 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
3996</div>
3997
Misha Brukman9d0919f2003-11-08 01:05:38 +00003998<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00003999
Misha Brukman9d0919f2003-11-08 01:05:38 +00004000<p>Variable argument support is defined in LLVM with the <a
Chris Lattnerfb6977d2006-01-13 23:26:01 +00004001 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner261efe92003-11-25 01:02:51 +00004002intrinsic functions. These functions are related to the similarly
4003named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004004
Chris Lattner261efe92003-11-25 01:02:51 +00004005<p>All of these functions operate on arguments that use a
4006target-specific value type "<tt>va_list</tt>". The LLVM assembly
4007language reference manual does not define what this type is, so all
Jeff Cohenb627eab2007-04-29 01:07:00 +00004008transformations should be prepared to handle these functions regardless of
4009the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004010
Chris Lattner374ab302006-05-15 17:26:46 +00004011<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +00004012instruction and the variable argument handling intrinsic functions are
4013used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004014
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004015<div class="doc_code">
Chris Lattner33aec9e2004-02-12 17:01:32 +00004016<pre>
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00004017define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +00004018 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +00004019 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +00004020 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00004021 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +00004022
4023 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +00004024 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +00004025
4026 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +00004027 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +00004028 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +00004029 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00004030 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +00004031
4032 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00004033 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +00004034 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +00004035}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00004036
4037declare void @llvm.va_start(i8*)
4038declare void @llvm.va_copy(i8*, i8*)
4039declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +00004040</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004041</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00004042
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004043</div>
4044
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004045<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00004046<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004047 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +00004048</div>
4049
4050
Misha Brukman9d0919f2003-11-08 01:05:38 +00004051<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004052<h5>Syntax:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +00004053<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004054<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00004055<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
4056<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4057href="#i_va_arg">va_arg</a></tt>.</p>
4058
4059<h5>Arguments:</h5>
4060
4061<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
4062
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004063<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00004064
4065<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
4066macro available in C. In a target-dependent way, it initializes the
Jeff Cohenb627eab2007-04-29 01:07:00 +00004067<tt>va_list</tt> element to which the argument points, so that the next call to
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00004068<tt>va_arg</tt> will produce the first variable argument passed to the function.
4069Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
Jeff Cohenb627eab2007-04-29 01:07:00 +00004070last argument of the function as the compiler can figure that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00004071
Misha Brukman9d0919f2003-11-08 01:05:38 +00004072</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00004073
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004074<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00004075<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004076 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +00004077</div>
4078
Misha Brukman9d0919f2003-11-08 01:05:38 +00004079<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004080<h5>Syntax:</h5>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004081<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004082<h5>Overview:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +00004083
Jeff Cohenb627eab2007-04-29 01:07:00 +00004084<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Reid Spencera3e435f2007-04-04 02:42:35 +00004085which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
Chris Lattner261efe92003-11-25 01:02:51 +00004086or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00004087
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004088<h5>Arguments:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +00004089
Jeff Cohenb627eab2007-04-29 01:07:00 +00004090<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00004091
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004092<h5>Semantics:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +00004093
Misha Brukman9d0919f2003-11-08 01:05:38 +00004094<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Jeff Cohenb627eab2007-04-29 01:07:00 +00004095macro available in C. In a target-dependent way, it destroys the
4096<tt>va_list</tt> element to which the argument points. Calls to <a
4097href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4098<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4099<tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00004100
Misha Brukman9d0919f2003-11-08 01:05:38 +00004101</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00004102
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004103<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +00004104<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004105 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +00004106</div>
4107
Misha Brukman9d0919f2003-11-08 01:05:38 +00004108<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +00004109
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004110<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00004111
4112<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004113 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00004114</pre>
4115
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004116<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00004117
Jeff Cohenb627eab2007-04-29 01:07:00 +00004118<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4119from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004120
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004121<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00004122
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00004123<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +00004124The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00004125
Chris Lattnerd7923912004-05-23 21:06:01 +00004126
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00004127<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00004128
Jeff Cohenb627eab2007-04-29 01:07:00 +00004129<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4130macro available in C. In a target-dependent way, it copies the source
4131<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4132intrinsic is necessary because the <tt><a href="#int_va_start">
4133llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4134example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004135
Misha Brukman9d0919f2003-11-08 01:05:38 +00004136</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00004137
Chris Lattner33aec9e2004-02-12 17:01:32 +00004138<!-- ======================================================================= -->
4139<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +00004140 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4141</div>
4142
4143<div class="doc_text">
4144
4145<p>
4146LLVM support for <a href="GarbageCollection.html">Accurate Garbage
4147Collection</a> requires the implementation and generation of these intrinsics.
Reid Spencera3e435f2007-04-04 02:42:35 +00004148These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
Chris Lattnerd7923912004-05-23 21:06:01 +00004149stack</a>, as well as garbage collector implementations that require <a
Reid Spencera3e435f2007-04-04 02:42:35 +00004150href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
Chris Lattnerd7923912004-05-23 21:06:01 +00004151Front-ends for type-safe garbage collected languages should generate these
4152intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4153href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4154</p>
Christopher Lamb303dae92007-12-17 01:00:21 +00004155
4156<p>The garbage collection intrinsics only operate on objects in the generic
4157 address space (address space zero).</p>
4158
Chris Lattnerd7923912004-05-23 21:06:01 +00004159</div>
4160
4161<!-- _______________________________________________________________________ -->
4162<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004163 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +00004164</div>
4165
4166<div class="doc_text">
4167
4168<h5>Syntax:</h5>
4169
4170<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00004171 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00004172</pre>
4173
4174<h5>Overview:</h5>
4175
John Criswell9e2485c2004-12-10 15:51:16 +00004176<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +00004177the code generator, and allows some metadata to be associated with it.</p>
4178
4179<h5>Arguments:</h5>
4180
4181<p>The first argument specifies the address of a stack object that contains the
4182root pointer. The second pointer (which must be either a constant or a global
4183value address) contains the meta-data to be associated with the root.</p>
4184
4185<h5>Semantics:</h5>
4186
4187<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
4188location. At compile-time, the code generator generates information to allow
Gordon Henriksene1433f22007-12-25 02:31:26 +00004189the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4190intrinsic may only be used in a function which <a href="#gc">specifies a GC
4191algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004192
4193</div>
4194
4195
4196<!-- _______________________________________________________________________ -->
4197<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004198 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +00004199</div>
4200
4201<div class="doc_text">
4202
4203<h5>Syntax:</h5>
4204
4205<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00004206 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00004207</pre>
4208
4209<h5>Overview:</h5>
4210
4211<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4212locations, allowing garbage collector implementations that require read
4213barriers.</p>
4214
4215<h5>Arguments:</h5>
4216
Chris Lattner80626e92006-03-14 20:02:51 +00004217<p>The second argument is the address to read from, which should be an address
4218allocated from the garbage collector. The first object is a pointer to the
4219start of the referenced object, if needed by the language runtime (otherwise
4220null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004221
4222<h5>Semantics:</h5>
4223
4224<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4225instruction, but may be replaced with substantially more complex code by the
Gordon Henriksene1433f22007-12-25 02:31:26 +00004226garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4227may only be used in a function which <a href="#gc">specifies a GC
4228algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004229
4230</div>
4231
4232
4233<!-- _______________________________________________________________________ -->
4234<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004235 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +00004236</div>
4237
4238<div class="doc_text">
4239
4240<h5>Syntax:</h5>
4241
4242<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00004243 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00004244</pre>
4245
4246<h5>Overview:</h5>
4247
4248<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4249locations, allowing garbage collector implementations that require write
4250barriers (such as generational or reference counting collectors).</p>
4251
4252<h5>Arguments:</h5>
4253
Chris Lattner80626e92006-03-14 20:02:51 +00004254<p>The first argument is the reference to store, the second is the start of the
4255object to store it to, and the third is the address of the field of Obj to
4256store to. If the runtime does not require a pointer to the object, Obj may be
4257null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004258
4259<h5>Semantics:</h5>
4260
4261<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4262instruction, but may be replaced with substantially more complex code by the
Gordon Henriksene1433f22007-12-25 02:31:26 +00004263garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4264may only be used in a function which <a href="#gc">specifies a GC
4265algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00004266
4267</div>
4268
4269
4270
4271<!-- ======================================================================= -->
4272<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +00004273 <a name="int_codegen">Code Generator Intrinsics</a>
4274</div>
4275
4276<div class="doc_text">
4277<p>
4278These intrinsics are provided by LLVM to expose special features that may only
4279be implemented with code generator support.
4280</p>
4281
4282</div>
4283
4284<!-- _______________________________________________________________________ -->
4285<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004286 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +00004287</div>
4288
4289<div class="doc_text">
4290
4291<h5>Syntax:</h5>
4292<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004293 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00004294</pre>
4295
4296<h5>Overview:</h5>
4297
4298<p>
Chris Lattner32b5d712006-10-15 20:05:59 +00004299The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4300target-specific value indicating the return address of the current function
4301or one of its callers.
Chris Lattner10610642004-02-14 04:08:35 +00004302</p>
4303
4304<h5>Arguments:</h5>
4305
4306<p>
4307The argument to this intrinsic indicates which function to return the address
4308for. Zero indicates the calling function, one indicates its caller, etc. The
4309argument is <b>required</b> to be a constant integer value.
4310</p>
4311
4312<h5>Semantics:</h5>
4313
4314<p>
4315The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4316the return address of the specified call frame, or zero if it cannot be
4317identified. The value returned by this intrinsic is likely to be incorrect or 0
4318for arguments other than zero, so it should only be used for debugging purposes.
4319</p>
4320
4321<p>
4322Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00004323aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00004324source-language caller.
4325</p>
4326</div>
4327
4328
4329<!-- _______________________________________________________________________ -->
4330<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004331 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +00004332</div>
4333
4334<div class="doc_text">
4335
4336<h5>Syntax:</h5>
4337<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00004338 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00004339</pre>
4340
4341<h5>Overview:</h5>
4342
4343<p>
Chris Lattner32b5d712006-10-15 20:05:59 +00004344The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4345target-specific frame pointer value for the specified stack frame.
Chris Lattner10610642004-02-14 04:08:35 +00004346</p>
4347
4348<h5>Arguments:</h5>
4349
4350<p>
4351The argument to this intrinsic indicates which function to return the frame
4352pointer for. Zero indicates the calling function, one indicates its caller,
4353etc. The argument is <b>required</b> to be a constant integer value.
4354</p>
4355
4356<h5>Semantics:</h5>
4357
4358<p>
4359The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4360the frame address of the specified call frame, or zero if it cannot be
4361identified. The value returned by this intrinsic is likely to be incorrect or 0
4362for arguments other than zero, so it should only be used for debugging purposes.
4363</p>
4364
4365<p>
4366Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +00004367aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +00004368source-language caller.
4369</p>
4370</div>
4371
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00004372<!-- _______________________________________________________________________ -->
4373<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004374 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +00004375</div>
4376
4377<div class="doc_text">
4378
4379<h5>Syntax:</h5>
4380<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00004381 declare i8 *@llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +00004382</pre>
4383
4384<h5>Overview:</h5>
4385
4386<p>
4387The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
Reid Spencera3e435f2007-04-04 02:42:35 +00004388the function stack, for use with <a href="#int_stackrestore">
Chris Lattner57e1f392006-01-13 02:03:13 +00004389<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4390features like scoped automatic variable sized arrays in C99.
4391</p>
4392
4393<h5>Semantics:</h5>
4394
4395<p>
4396This intrinsic returns a opaque pointer value that can be passed to <a
Reid Spencera3e435f2007-04-04 02:42:35 +00004397href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
Chris Lattner57e1f392006-01-13 02:03:13 +00004398<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4399<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4400state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4401practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4402that were allocated after the <tt>llvm.stacksave</tt> was executed.
4403</p>
4404
4405</div>
4406
4407<!-- _______________________________________________________________________ -->
4408<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004409 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +00004410</div>
4411
4412<div class="doc_text">
4413
4414<h5>Syntax:</h5>
4415<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004416 declare void @llvm.stackrestore(i8 * %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +00004417</pre>
4418
4419<h5>Overview:</h5>
4420
4421<p>
4422The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4423the function stack to the state it was in when the corresponding <a
Reid Spencera3e435f2007-04-04 02:42:35 +00004424href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
Chris Lattner57e1f392006-01-13 02:03:13 +00004425useful for implementing language features like scoped automatic variable sized
4426arrays in C99.
4427</p>
4428
4429<h5>Semantics:</h5>
4430
4431<p>
Reid Spencera3e435f2007-04-04 02:42:35 +00004432See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
Chris Lattner57e1f392006-01-13 02:03:13 +00004433</p>
4434
4435</div>
4436
4437
4438<!-- _______________________________________________________________________ -->
4439<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004440 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00004441</div>
4442
4443<div class="doc_text">
4444
4445<h5>Syntax:</h5>
4446<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00004447 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00004448</pre>
4449
4450<h5>Overview:</h5>
4451
4452
4453<p>
4454The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +00004455a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4456no
4457effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +00004458characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00004459</p>
4460
4461<h5>Arguments:</h5>
4462
4463<p>
4464<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4465determining if the fetch should be for a read (0) or write (1), and
4466<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +00004467locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00004468<tt>locality</tt> arguments must be constant integers.
4469</p>
4470
4471<h5>Semantics:</h5>
4472
4473<p>
4474This intrinsic does not modify the behavior of the program. In particular,
4475prefetches cannot trap and do not produce a value. On targets that support this
4476intrinsic, the prefetch can provide hints to the processor cache for better
4477performance.
4478</p>
4479
4480</div>
4481
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00004482<!-- _______________________________________________________________________ -->
4483<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004484 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00004485</div>
4486
4487<div class="doc_text">
4488
4489<h5>Syntax:</h5>
4490<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00004491 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00004492</pre>
4493
4494<h5>Overview:</h5>
4495
4496
4497<p>
John Criswellfc6b8952005-05-16 16:17:45 +00004498The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
4499(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00004500code to simulators and other tools. The method is target specific, but it is
4501expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +00004502The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +00004503after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb3e7afd2006-03-24 07:16:10 +00004504optimizations. The intended use is to be inserted after optimizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +00004505correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00004506</p>
4507
4508<h5>Arguments:</h5>
4509
4510<p>
4511<tt>id</tt> is a numerical id identifying the marker.
4512</p>
4513
4514<h5>Semantics:</h5>
4515
4516<p>
4517This intrinsic does not modify the behavior of the program. Backends that do not
4518support this intrinisic may ignore it.
4519</p>
4520
4521</div>
4522
Andrew Lenharth51b8d542005-11-11 16:47:30 +00004523<!-- _______________________________________________________________________ -->
4524<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004525 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00004526</div>
4527
4528<div class="doc_text">
4529
4530<h5>Syntax:</h5>
4531<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004532 declare i64 @llvm.readcyclecounter( )
Andrew Lenharth51b8d542005-11-11 16:47:30 +00004533</pre>
4534
4535<h5>Overview:</h5>
4536
4537
4538<p>
4539The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
4540counter register (or similar low latency, high accuracy clocks) on those targets
4541that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
4542As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
4543should only be used for small timings.
4544</p>
4545
4546<h5>Semantics:</h5>
4547
4548<p>
4549When directly supported, reading the cycle counter should not modify any memory.
4550Implementations are allowed to either return a application specific value or a
4551system wide value. On backends without support, this is lowered to a constant 0.
4552</p>
4553
4554</div>
4555
Chris Lattner10610642004-02-14 04:08:35 +00004556<!-- ======================================================================= -->
4557<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +00004558 <a name="int_libc">Standard C Library Intrinsics</a>
4559</div>
4560
4561<div class="doc_text">
4562<p>
Chris Lattner10610642004-02-14 04:08:35 +00004563LLVM provides intrinsics for a few important standard C library functions.
4564These intrinsics allow source-language front-ends to pass information about the
4565alignment of the pointer arguments to the code generator, providing opportunity
4566for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +00004567</p>
4568
4569</div>
4570
4571<!-- _______________________________________________________________________ -->
4572<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004573 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattner33aec9e2004-02-12 17:01:32 +00004574</div>
4575
4576<div class="doc_text">
4577
4578<h5>Syntax:</h5>
4579<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004580 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +00004581 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004582 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +00004583 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00004584</pre>
4585
4586<h5>Overview:</h5>
4587
4588<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004589The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00004590location to the destination location.
4591</p>
4592
4593<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004594Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4595intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner33aec9e2004-02-12 17:01:32 +00004596</p>
4597
4598<h5>Arguments:</h5>
4599
4600<p>
4601The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00004602the source. The third argument is an integer argument
Chris Lattner33aec9e2004-02-12 17:01:32 +00004603specifying the number of bytes to copy, and the fourth argument is the alignment
4604of the source and destination locations.
4605</p>
4606
Chris Lattner3301ced2004-02-12 21:18:15 +00004607<p>
4608If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00004609the caller guarantees that both the source and destination pointers are aligned
4610to that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00004611</p>
4612
Chris Lattner33aec9e2004-02-12 17:01:32 +00004613<h5>Semantics:</h5>
4614
4615<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004616The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +00004617location to the destination location, which are not allowed to overlap. It
4618copies "len" bytes of memory over. If the argument is known to be aligned to
4619some boundary, this can be specified as the fourth argument, otherwise it should
4620be set to 0 or 1.
4621</p>
4622</div>
4623
4624
Chris Lattner0eb51b42004-02-12 18:10:10 +00004625<!-- _______________________________________________________________________ -->
4626<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004627 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattner0eb51b42004-02-12 18:10:10 +00004628</div>
4629
4630<div class="doc_text">
4631
4632<h5>Syntax:</h5>
4633<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004634 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +00004635 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004636 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +00004637 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00004638</pre>
4639
4640<h5>Overview:</h5>
4641
4642<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004643The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4644location to the destination location. It is similar to the
Chris Lattner4b2cbcf2008-01-06 19:51:52 +00004645'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattner0eb51b42004-02-12 18:10:10 +00004646</p>
4647
4648<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004649Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4650intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner0eb51b42004-02-12 18:10:10 +00004651</p>
4652
4653<h5>Arguments:</h5>
4654
4655<p>
4656The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +00004657the source. The third argument is an integer argument
Chris Lattner0eb51b42004-02-12 18:10:10 +00004658specifying the number of bytes to copy, and the fourth argument is the alignment
4659of the source and destination locations.
4660</p>
4661
Chris Lattner3301ced2004-02-12 21:18:15 +00004662<p>
4663If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00004664the caller guarantees that the source and destination pointers are aligned to
4665that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +00004666</p>
4667
Chris Lattner0eb51b42004-02-12 18:10:10 +00004668<h5>Semantics:</h5>
4669
4670<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004671The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner0eb51b42004-02-12 18:10:10 +00004672location to the destination location, which may overlap. It
4673copies "len" bytes of memory over. If the argument is known to be aligned to
4674some boundary, this can be specified as the fourth argument, otherwise it should
4675be set to 0 or 1.
4676</p>
4677</div>
4678
Chris Lattner8ff75902004-01-06 05:31:32 +00004679
Chris Lattner10610642004-02-14 04:08:35 +00004680<!-- _______________________________________________________________________ -->
4681<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004682 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner10610642004-02-14 04:08:35 +00004683</div>
4684
4685<div class="doc_text">
4686
4687<h5>Syntax:</h5>
4688<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004689 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +00004690 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00004691 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +00004692 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00004693</pre>
4694
4695<h5>Overview:</h5>
4696
4697<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004698The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner10610642004-02-14 04:08:35 +00004699byte value.
4700</p>
4701
4702<p>
4703Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4704does not return a value, and takes an extra alignment argument.
4705</p>
4706
4707<h5>Arguments:</h5>
4708
4709<p>
4710The first argument is a pointer to the destination to fill, the second is the
Chris Lattner5b310c32006-03-03 00:07:20 +00004711byte value to fill it with, the third argument is an integer
Chris Lattner10610642004-02-14 04:08:35 +00004712argument specifying the number of bytes to fill, and the fourth argument is the
4713known alignment of destination location.
4714</p>
4715
4716<p>
4717If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +00004718the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner10610642004-02-14 04:08:35 +00004719</p>
4720
4721<h5>Semantics:</h5>
4722
4723<p>
Chris Lattner5b310c32006-03-03 00:07:20 +00004724The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4725the
Chris Lattner10610642004-02-14 04:08:35 +00004726destination location. If the argument is known to be aligned to some boundary,
4727this can be specified as the fourth argument, otherwise it should be set to 0 or
47281.
4729</p>
4730</div>
4731
4732
Chris Lattner32006282004-06-11 02:28:03 +00004733<!-- _______________________________________________________________________ -->
4734<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004735 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +00004736</div>
4737
4738<div class="doc_text">
4739
4740<h5>Syntax:</h5>
Dale Johannesen408f9c12007-10-02 17:47:38 +00004741<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
Dan Gohman91c284c2007-10-15 20:30:11 +00004742floating point or vector of floating point type. Not all targets support all
4743types however.
Chris Lattnera4d74142005-07-21 01:29:16 +00004744<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +00004745 declare float @llvm.sqrt.f32(float %Val)
4746 declare double @llvm.sqrt.f64(double %Val)
4747 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
4748 declare fp128 @llvm.sqrt.f128(fp128 %Val)
4749 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +00004750</pre>
4751
4752<h5>Overview:</h5>
4753
4754<p>
Reid Spencer0b118202006-01-16 21:12:35 +00004755The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman91c284c2007-10-15 20:30:11 +00004756returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Chris Lattnera4d74142005-07-21 01:29:16 +00004757<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattner103e2d72008-01-29 07:00:44 +00004758negative numbers other than -0.0 (which allows for better optimization, because
4759there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
4760defined to return -0.0 like IEEE sqrt.
Chris Lattnera4d74142005-07-21 01:29:16 +00004761</p>
4762
4763<h5>Arguments:</h5>
4764
4765<p>
4766The argument and return value are floating point numbers of the same type.
4767</p>
4768
4769<h5>Semantics:</h5>
4770
4771<p>
Dan Gohmand6257fe2007-07-16 14:37:41 +00004772This function returns the sqrt of the specified operand if it is a nonnegative
Chris Lattnera4d74142005-07-21 01:29:16 +00004773floating point number.
4774</p>
4775</div>
4776
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004777<!-- _______________________________________________________________________ -->
4778<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004779 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004780</div>
4781
4782<div class="doc_text">
4783
4784<h5>Syntax:</h5>
Dale Johannesen408f9c12007-10-02 17:47:38 +00004785<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
Dan Gohman91c284c2007-10-15 20:30:11 +00004786floating point or vector of floating point type. Not all targets support all
4787types however.
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004788<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +00004789 declare float @llvm.powi.f32(float %Val, i32 %power)
4790 declare double @llvm.powi.f64(double %Val, i32 %power)
4791 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
4792 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
4793 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004794</pre>
4795
4796<h5>Overview:</h5>
4797
4798<p>
4799The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4800specified (positive or negative) power. The order of evaluation of
Dan Gohman91c284c2007-10-15 20:30:11 +00004801multiplications is not defined. When a vector of floating point type is
4802used, the second argument remains a scalar integer value.
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004803</p>
4804
4805<h5>Arguments:</h5>
4806
4807<p>
4808The second argument is an integer power, and the first is a value to raise to
4809that power.
4810</p>
4811
4812<h5>Semantics:</h5>
4813
4814<p>
4815This function returns the first value raised to the second power with an
4816unspecified sequence of rounding operations.</p>
4817</div>
4818
Dan Gohman91c284c2007-10-15 20:30:11 +00004819<!-- _______________________________________________________________________ -->
4820<div class="doc_subsubsection">
4821 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
4822</div>
4823
4824<div class="doc_text">
4825
4826<h5>Syntax:</h5>
4827<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
4828floating point or vector of floating point type. Not all targets support all
4829types however.
4830<pre>
4831 declare float @llvm.sin.f32(float %Val)
4832 declare double @llvm.sin.f64(double %Val)
4833 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
4834 declare fp128 @llvm.sin.f128(fp128 %Val)
4835 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
4836</pre>
4837
4838<h5>Overview:</h5>
4839
4840<p>
4841The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
4842</p>
4843
4844<h5>Arguments:</h5>
4845
4846<p>
4847The argument and return value are floating point numbers of the same type.
4848</p>
4849
4850<h5>Semantics:</h5>
4851
4852<p>
4853This function returns the sine of the specified operand, returning the
4854same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmanba83b7e2007-10-17 18:05:13 +00004855conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00004856</div>
4857
4858<!-- _______________________________________________________________________ -->
4859<div class="doc_subsubsection">
4860 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
4861</div>
4862
4863<div class="doc_text">
4864
4865<h5>Syntax:</h5>
4866<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
4867floating point or vector of floating point type. Not all targets support all
4868types however.
4869<pre>
4870 declare float @llvm.cos.f32(float %Val)
4871 declare double @llvm.cos.f64(double %Val)
4872 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
4873 declare fp128 @llvm.cos.f128(fp128 %Val)
4874 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
4875</pre>
4876
4877<h5>Overview:</h5>
4878
4879<p>
4880The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
4881</p>
4882
4883<h5>Arguments:</h5>
4884
4885<p>
4886The argument and return value are floating point numbers of the same type.
4887</p>
4888
4889<h5>Semantics:</h5>
4890
4891<p>
4892This function returns the cosine of the specified operand, returning the
4893same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmanba83b7e2007-10-17 18:05:13 +00004894conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00004895</div>
4896
4897<!-- _______________________________________________________________________ -->
4898<div class="doc_subsubsection">
4899 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
4900</div>
4901
4902<div class="doc_text">
4903
4904<h5>Syntax:</h5>
4905<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
4906floating point or vector of floating point type. Not all targets support all
4907types however.
4908<pre>
4909 declare float @llvm.pow.f32(float %Val, float %Power)
4910 declare double @llvm.pow.f64(double %Val, double %Power)
4911 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
4912 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
4913 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
4914</pre>
4915
4916<h5>Overview:</h5>
4917
4918<p>
4919The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
4920specified (positive or negative) power.
4921</p>
4922
4923<h5>Arguments:</h5>
4924
4925<p>
4926The second argument is a floating point power, and the first is a value to
4927raise to that power.
4928</p>
4929
4930<h5>Semantics:</h5>
4931
4932<p>
4933This function returns the first value raised to the second power,
4934returning the
4935same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmanba83b7e2007-10-17 18:05:13 +00004936conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00004937</div>
4938
Chris Lattnerf4d252d2006-09-08 06:34:02 +00004939
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004940<!-- ======================================================================= -->
4941<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +00004942 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004943</div>
4944
4945<div class="doc_text">
4946<p>
Nate Begeman7e36c472006-01-13 23:26:38 +00004947LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004948These allow efficient code generation for some algorithms.
4949</p>
4950
4951</div>
4952
4953<!-- _______________________________________________________________________ -->
4954<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +00004955 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman7e36c472006-01-13 23:26:38 +00004956</div>
4957
4958<div class="doc_text">
4959
4960<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00004961<p>This is an overloaded intrinsic function. You can use bswap on any integer
Chandler Carruth69940402007-08-04 01:51:18 +00004962type that is an even number of bytes (i.e. BitWidth % 16 == 0).
Nate Begeman7e36c472006-01-13 23:26:38 +00004963<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00004964 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
4965 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
4966 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +00004967</pre>
4968
4969<h5>Overview:</h5>
4970
4971<p>
Reid Spencer338ea092007-04-02 02:25:19 +00004972The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
Reid Spencer409e28f2007-04-01 08:04:23 +00004973values with an even number of bytes (positive multiple of 16 bits). These are
4974useful for performing operations on data that is not in the target's native
4975byte order.
Nate Begeman7e36c472006-01-13 23:26:38 +00004976</p>
4977
4978<h5>Semantics:</h5>
4979
4980<p>
Chandler Carruth69940402007-08-04 01:51:18 +00004981The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Reid Spencerca86e162006-12-31 07:07:53 +00004982and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
4983intrinsic returns an i32 value that has the four bytes of the input i32
4984swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the returned
Chandler Carruth69940402007-08-04 01:51:18 +00004985i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
4986<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Reid Spencer409e28f2007-04-01 08:04:23 +00004987additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
Nate Begeman7e36c472006-01-13 23:26:38 +00004988</p>
4989
4990</div>
4991
4992<!-- _______________________________________________________________________ -->
4993<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +00004994 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00004995</div>
4996
4997<div class="doc_text">
4998
4999<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00005000<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
5001width. Not all targets support all bit widths however.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005002<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00005003 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5004 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00005005 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00005006 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5007 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005008</pre>
5009
5010<h5>Overview:</h5>
5011
5012<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +00005013The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5014value.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005015</p>
5016
5017<h5>Arguments:</h5>
5018
5019<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00005020The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +00005021integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005022</p>
5023
5024<h5>Semantics:</h5>
5025
5026<p>
5027The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5028</p>
5029</div>
5030
5031<!-- _______________________________________________________________________ -->
5032<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00005033 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005034</div>
5035
5036<div class="doc_text">
5037
5038<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00005039<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
5040integer bit width. Not all targets support all bit widths however.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005041<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00005042 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5043 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00005044 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00005045 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5046 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005047</pre>
5048
5049<h5>Overview:</h5>
5050
5051<p>
Reid Spencer0b118202006-01-16 21:12:35 +00005052The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5053leading zeros in a variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005054</p>
5055
5056<h5>Arguments:</h5>
5057
5058<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +00005059The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +00005060integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005061</p>
5062
5063<h5>Semantics:</h5>
5064
5065<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00005066The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5067in a variable. If the src == 0 then the result is the size in bits of the type
Reid Spencerca86e162006-12-31 07:07:53 +00005068of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +00005069</p>
5070</div>
Chris Lattner32006282004-06-11 02:28:03 +00005071
5072
Chris Lattnereff29ab2005-05-15 19:39:26 +00005073
5074<!-- _______________________________________________________________________ -->
5075<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +00005076 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +00005077</div>
5078
5079<div class="doc_text">
5080
5081<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00005082<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
5083integer bit width. Not all targets support all bit widths however.
Chris Lattnereff29ab2005-05-15 19:39:26 +00005084<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00005085 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5086 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00005087 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00005088 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5089 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +00005090</pre>
5091
5092<h5>Overview:</h5>
5093
5094<p>
Reid Spencer0b118202006-01-16 21:12:35 +00005095The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5096trailing zeros.
Chris Lattnereff29ab2005-05-15 19:39:26 +00005097</p>
5098
5099<h5>Arguments:</h5>
5100
5101<p>
5102The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +00005103integer type. The return type must match the argument type.
Chris Lattnereff29ab2005-05-15 19:39:26 +00005104</p>
5105
5106<h5>Semantics:</h5>
5107
5108<p>
5109The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5110in a variable. If the src == 0 then the result is the size in bits of the type
5111of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5112</p>
5113</div>
5114
Reid Spencer497d93e2007-04-01 08:27:01 +00005115<!-- _______________________________________________________________________ -->
5116<div class="doc_subsubsection">
Reid Spencerbeacf662007-04-10 02:51:31 +00005117 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
Reid Spencera13ba7d2007-04-01 19:00:37 +00005118</div>
5119
5120<div class="doc_text">
5121
5122<h5>Syntax:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +00005123<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Reid Spencera13ba7d2007-04-01 19:00:37 +00005124on any integer bit width.
5125<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00005126 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5127 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Reid Spencera13ba7d2007-04-01 19:00:37 +00005128</pre>
5129
5130<h5>Overview:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +00005131<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
Reid Spencera13ba7d2007-04-01 19:00:37 +00005132range of bits from an integer value and returns them in the same bit width as
5133the original value.</p>
5134
5135<h5>Arguments:</h5>
5136<p>The first argument, <tt>%val</tt> and the result may be integer types of
5137any bit width but they must have the same bit width. The second and third
Reid Spencera3e435f2007-04-04 02:42:35 +00005138arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
Reid Spencera13ba7d2007-04-01 19:00:37 +00005139
5140<h5>Semantics:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +00005141<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
Reid Spencera3e435f2007-04-04 02:42:35 +00005142of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5143<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5144operates in forward mode.</p>
5145<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5146right by <tt>%loBit</tt> bits and then ANDing it with a mask with
Reid Spencera13ba7d2007-04-01 19:00:37 +00005147only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5148<ol>
5149 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5150 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5151 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5152 to determine the number of bits to retain.</li>
5153 <li>A mask of the retained bits is created by shifting a -1 value.</li>
5154 <li>The mask is ANDed with <tt>%val</tt> to produce the result.
5155</ol>
Reid Spencerd6a85b52007-05-14 16:14:57 +00005156<p>In reverse mode, a similar computation is made except that the bits are
5157returned in the reverse order. So, for example, if <tt>X</tt> has the value
5158<tt>i16 0x0ACF (101011001111)</tt> and we apply
5159<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5160<tt>i16 0x0026 (000000100110)</tt>.</p>
Reid Spencera13ba7d2007-04-01 19:00:37 +00005161</div>
5162
Reid Spencerf86037f2007-04-11 23:23:49 +00005163<div class="doc_subsubsection">
5164 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5165</div>
5166
5167<div class="doc_text">
5168
5169<h5>Syntax:</h5>
5170<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
5171on any integer bit width.
5172<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00005173 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5174 declare i29 @llvm.part.set.i29.i9 (i29 %val, i9 %repl, i32 %lo, i32 %hi)
Reid Spencerf86037f2007-04-11 23:23:49 +00005175</pre>
5176
5177<h5>Overview:</h5>
5178<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5179of bits in an integer value with another integer value. It returns the integer
5180with the replaced bits.</p>
5181
5182<h5>Arguments:</h5>
5183<p>The first argument, <tt>%val</tt> and the result may be integer types of
5184any bit width but they must have the same bit width. <tt>%val</tt> is the value
5185whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5186integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5187type since they specify only a bit index.</p>
5188
5189<h5>Semantics:</h5>
5190<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5191of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5192<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5193operates in forward mode.</p>
5194<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5195truncating it down to the size of the replacement area or zero extending it
5196up to that size.</p>
5197<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5198are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5199in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
5200to the <tt>%hi</tt>th bit.
Reid Spencerc6749c42007-05-14 16:50:20 +00005201<p>In reverse mode, a similar computation is made except that the bits are
5202reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
5203<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.
Reid Spencerf86037f2007-04-11 23:23:49 +00005204<h5>Examples:</h5>
5205<pre>
Reid Spencerf0dbf642007-04-12 01:03:03 +00005206 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
Reid Spencerc6749c42007-05-14 16:50:20 +00005207 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5208 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5209 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
Reid Spencerf0dbf642007-04-12 01:03:03 +00005210 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
Reid Spencerc8910842007-04-11 23:49:50 +00005211</pre>
Reid Spencerf86037f2007-04-11 23:23:49 +00005212</div>
5213
Chris Lattner8ff75902004-01-06 05:31:32 +00005214<!-- ======================================================================= -->
5215<div class="doc_subsection">
5216 <a name="int_debugger">Debugger Intrinsics</a>
5217</div>
5218
5219<div class="doc_text">
5220<p>
5221The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5222are described in the <a
5223href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5224Debugging</a> document.
5225</p>
5226</div>
5227
5228
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00005229<!-- ======================================================================= -->
5230<div class="doc_subsection">
5231 <a name="int_eh">Exception Handling Intrinsics</a>
5232</div>
5233
5234<div class="doc_text">
5235<p> The LLVM exception handling intrinsics (which all start with
5236<tt>llvm.eh.</tt> prefix), are described in the <a
5237href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5238Handling</a> document. </p>
5239</div>
5240
Tanya Lattner6d806e92007-06-15 20:50:54 +00005241<!-- ======================================================================= -->
5242<div class="doc_subsection">
Duncan Sandsf7331b32007-09-11 14:10:23 +00005243 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +00005244</div>
5245
5246<div class="doc_text">
5247<p>
Duncan Sandsf7331b32007-09-11 14:10:23 +00005248 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands36397f52007-07-27 12:58:54 +00005249 the <tt>nest</tt> attribute, from a function. The result is a callable
5250 function pointer lacking the nest parameter - the caller does not need
5251 to provide a value for it. Instead, the value to use is stored in
5252 advance in a "trampoline", a block of memory usually allocated
5253 on the stack, which also contains code to splice the nest value into the
5254 argument list. This is used to implement the GCC nested function address
5255 extension.
5256</p>
5257<p>
5258 For example, if the function is
5259 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
Bill Wendling03295ca2007-09-22 09:23:55 +00005260 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as follows:</p>
Duncan Sands36397f52007-07-27 12:58:54 +00005261<pre>
Duncan Sandsf7331b32007-09-11 14:10:23 +00005262 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5263 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5264 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5265 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +00005266</pre>
Bill Wendling03295ca2007-09-22 09:23:55 +00005267 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5268 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands36397f52007-07-27 12:58:54 +00005269</div>
5270
5271<!-- _______________________________________________________________________ -->
5272<div class="doc_subsubsection">
5273 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5274</div>
5275<div class="doc_text">
5276<h5>Syntax:</h5>
5277<pre>
Duncan Sandsf7331b32007-09-11 14:10:23 +00005278declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands36397f52007-07-27 12:58:54 +00005279</pre>
5280<h5>Overview:</h5>
5281<p>
Duncan Sandsf7331b32007-09-11 14:10:23 +00005282 This fills the memory pointed to by <tt>tramp</tt> with code
5283 and returns a function pointer suitable for executing it.
Duncan Sands36397f52007-07-27 12:58:54 +00005284</p>
5285<h5>Arguments:</h5>
5286<p>
5287 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5288 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5289 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sandsc00c2ba2007-08-22 23:39:54 +00005290 intrinsic. Note that the size and the alignment are target-specific - LLVM
5291 currently provides no portable way of determining them, so a front-end that
5292 generates this intrinsic needs to have some target-specific knowledge.
5293 The <tt>func</tt> argument must hold a function bitcast to an <tt>i8*</tt>.
Duncan Sands36397f52007-07-27 12:58:54 +00005294</p>
5295<h5>Semantics:</h5>
5296<p>
5297 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sandsf7331b32007-09-11 14:10:23 +00005298 dependent code, turning it into a function. A pointer to this function is
5299 returned, but needs to be bitcast to an
Duncan Sands36397f52007-07-27 12:58:54 +00005300 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sandsf7331b32007-09-11 14:10:23 +00005301 before being called. The new function's signature is the same as that of
5302 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5303 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5304 of pointer type. Calling the new function is equivalent to calling
5305 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5306 missing <tt>nest</tt> argument. If, after calling
5307 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5308 modified, then the effect of any later call to the returned function pointer is
5309 undefined.
Duncan Sands36397f52007-07-27 12:58:54 +00005310</p>
5311</div>
5312
5313<!-- ======================================================================= -->
5314<div class="doc_subsection">
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00005315 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5316</div>
5317
5318<div class="doc_text">
5319<p>
5320 These intrinsic functions expand the "universal IR" of LLVM to represent
5321 hardware constructs for atomic operations and memory synchronization. This
5322 provides an interface to the hardware, not an interface to the programmer. It
5323 is aimed at a low enough level to allow any programming models or APIs which
5324 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5325 hardware behavior. Just as hardware provides a "universal IR" for source
5326 languages, it also provides a starting point for developing a "universal"
5327 atomic operation and synchronization IR.
5328</p>
5329<p>
5330 These do <em>not</em> form an API such as high-level threading libraries,
5331 software transaction memory systems, atomic primitives, and intrinsic
5332 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5333 application libraries. The hardware interface provided by LLVM should allow
5334 a clean implementation of all of these APIs and parallel programming models.
5335 No one model or paradigm should be selected above others unless the hardware
5336 itself ubiquitously does so.
5337
5338</p>
5339</div>
5340
5341<!-- _______________________________________________________________________ -->
5342<div class="doc_subsubsection">
5343 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5344</div>
5345<div class="doc_text">
5346<h5>Syntax:</h5>
5347<pre>
5348declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5349i1 &lt;device&gt; )
5350
5351</pre>
5352<h5>Overview:</h5>
5353<p>
5354 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5355 specific pairs of memory access types.
5356</p>
5357<h5>Arguments:</h5>
5358<p>
5359 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5360 The first four arguments enables a specific barrier as listed below. The fith
5361 argument specifies that the barrier applies to io or device or uncached memory.
5362
5363</p>
5364 <ul>
5365 <li><tt>ll</tt>: load-load barrier</li>
5366 <li><tt>ls</tt>: load-store barrier</li>
5367 <li><tt>sl</tt>: store-load barrier</li>
5368 <li><tt>ss</tt>: store-store barrier</li>
5369 <li><tt>device</tt>: barrier applies to device and uncached memory also.
5370 </ul>
5371<h5>Semantics:</h5>
5372<p>
5373 This intrinsic causes the system to enforce some ordering constraints upon
5374 the loads and stores of the program. This barrier does not indicate
5375 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5376 which they occur. For any of the specified pairs of load and store operations
5377 (f.ex. load-load, or store-load), all of the first operations preceding the
5378 barrier will complete before any of the second operations succeeding the
5379 barrier begin. Specifically the semantics for each pairing is as follows:
5380</p>
5381 <ul>
5382 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5383 after the barrier begins.</li>
5384
5385 <li><tt>ls</tt>: All loads before the barrier must complete before any
5386 store after the barrier begins.</li>
5387 <li><tt>ss</tt>: All stores before the barrier must complete before any
5388 store after the barrier begins.</li>
5389 <li><tt>sl</tt>: All stores before the barrier must complete before any
5390 load after the barrier begins.</li>
5391 </ul>
5392<p>
5393 These semantics are applied with a logical "and" behavior when more than one
5394 is enabled in a single memory barrier intrinsic.
5395</p>
5396<p>
5397 Backends may implement stronger barriers than those requested when they do not
5398 support as fine grained a barrier as requested. Some architectures do not
5399 need all types of barriers and on such architectures, these become noops.
5400</p>
5401<h5>Example:</h5>
5402<pre>
5403%ptr = malloc i32
5404 store i32 4, %ptr
5405
5406%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5407 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5408 <i>; guarantee the above finishes</i>
5409 store i32 8, %ptr <i>; before this begins</i>
5410</pre>
5411</div>
5412
Andrew Lenharthab0b9492008-02-21 06:45:13 +00005413<!-- _______________________________________________________________________ -->
5414<div class="doc_subsubsection">
5415 <a name="int_atomic_lcs">'<tt>llvm.atomic.lcs.*</tt>' Intrinsic</a>
5416</div>
5417<div class="doc_text">
5418<h5>Syntax:</h5>
5419<p>
5420 This is an overloaded intrinsic. You can use <tt>llvm.atomic.lcs</tt> on any
5421 integer bit width. Not all targets support all bit widths however.</p>
5422
5423<pre>
5424declare i8 @llvm.atomic.lcs.i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5425declare i16 @llvm.atomic.lcs.i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5426declare i32 @llvm.atomic.lcs.i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5427declare i64 @llvm.atomic.lcs.i64( i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt; )
5428
5429</pre>
5430<h5>Overview:</h5>
5431<p>
5432 This loads a value in memory and compares it to a given value. If they are
5433 equal, it stores a new value into the memory.
5434</p>
5435<h5>Arguments:</h5>
5436<p>
5437 The <tt>llvm.atomic.lcs</tt> intrinsic takes three arguments. The result as
5438 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5439 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5440 this integer type. While any bit width integer may be used, targets may only
5441 lower representations they support in hardware.
5442
5443</p>
5444<h5>Semantics:</h5>
5445<p>
5446 This entire intrinsic must be executed atomically. It first loads the value
5447 in memory pointed to by <tt>ptr</tt> and compares it with the value
5448 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5449 loaded value is yielded in all cases. This provides the equivalent of an
5450 atomic compare-and-swap operation within the SSA framework.
5451</p>
5452<h5>Examples:</h5>
5453
5454<pre>
5455%ptr = malloc i32
5456 store i32 4, %ptr
5457
5458%val1 = add i32 4, 4
5459%result1 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 4, %val1 )
5460 <i>; yields {i32}:result1 = 4</i>
5461%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5462%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5463
5464%val2 = add i32 1, 1
5465%result2 = call i32 @llvm.atomic.lcs.i32( i32* %ptr, i32 5, %val2 )
5466 <i>; yields {i32}:result2 = 8</i>
5467%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
5468
5469%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
5470</pre>
5471</div>
5472
5473<!-- _______________________________________________________________________ -->
5474<div class="doc_subsubsection">
5475 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
5476</div>
5477<div class="doc_text">
5478<h5>Syntax:</h5>
5479
5480<p>
5481 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
5482 integer bit width. Not all targets support all bit widths however.</p>
5483<pre>
5484declare i8 @llvm.atomic.swap.i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
5485declare i16 @llvm.atomic.swap.i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
5486declare i32 @llvm.atomic.swap.i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
5487declare i64 @llvm.atomic.swap.i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
5488
5489</pre>
5490<h5>Overview:</h5>
5491<p>
5492 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
5493 the value from memory. It then stores the value in <tt>val</tt> in the memory
5494 at <tt>ptr</tt>.
5495</p>
5496<h5>Arguments:</h5>
5497
5498<p>
5499 The <tt>llvm.atomic.ls</tt> intrinsic takes two arguments. Both the
5500 <tt>val</tt> argument and the result must be integers of the same bit width.
5501 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
5502 integer type. The targets may only lower integer representations they
5503 support.
5504</p>
5505<h5>Semantics:</h5>
5506<p>
5507 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
5508 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
5509 equivalent of an atomic swap operation within the SSA framework.
5510
5511</p>
5512<h5>Examples:</h5>
5513<pre>
5514%ptr = malloc i32
5515 store i32 4, %ptr
5516
5517%val1 = add i32 4, 4
5518%result1 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val1 )
5519 <i>; yields {i32}:result1 = 4</i>
5520%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5521%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5522
5523%val2 = add i32 1, 1
5524%result2 = call i32 @llvm.atomic.swap.i32( i32* %ptr, i32 %val2 )
5525 <i>; yields {i32}:result2 = 8</i>
5526
5527%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
5528%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
5529</pre>
5530</div>
5531
5532<!-- _______________________________________________________________________ -->
5533<div class="doc_subsubsection">
5534 <a name="int_atomic_las">'<tt>llvm.atomic.las.*</tt>' Intrinsic</a>
5535
5536</div>
5537<div class="doc_text">
5538<h5>Syntax:</h5>
5539<p>
5540 This is an overloaded intrinsic. You can use <tt>llvm.atomic.las</tt> on any
5541 integer bit width. Not all targets support all bit widths however.</p>
5542<pre>
5543declare i8 @llvm.atomic.las.i8.( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
5544declare i16 @llvm.atomic.las.i16.( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
5545declare i32 @llvm.atomic.las.i32.( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
5546declare i64 @llvm.atomic.las.i64.( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
5547
5548</pre>
5549<h5>Overview:</h5>
5550<p>
5551 This intrinsic adds <tt>delta</tt> to the value stored in memory at
5552 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
5553</p>
5554<h5>Arguments:</h5>
5555<p>
5556
5557 The intrinsic takes two arguments, the first a pointer to an integer value
5558 and the second an integer value. The result is also an integer value. These
5559 integer types can have any bit width, but they must all have the same bit
5560 width. The targets may only lower integer representations they support.
5561</p>
5562<h5>Semantics:</h5>
5563<p>
5564 This intrinsic does a series of operations atomically. It first loads the
5565 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
5566 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
5567</p>
5568
5569<h5>Examples:</h5>
5570<pre>
5571%ptr = malloc i32
5572 store i32 4, %ptr
5573%result1 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 4 )
5574 <i>; yields {i32}:result1 = 4</i>
5575%result2 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 2 )
5576 <i>; yields {i32}:result2 = 8</i>
5577%result3 = call i32 @llvm.atomic.las.i32( i32* %ptr, i32 5 )
5578 <i>; yields {i32}:result3 = 10</i>
5579%memval = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
5580</pre>
5581</div>
5582
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00005583
5584<!-- ======================================================================= -->
5585<div class="doc_subsection">
Tanya Lattner6d806e92007-06-15 20:50:54 +00005586 <a name="int_general">General Intrinsics</a>
5587</div>
5588
5589<div class="doc_text">
5590<p> This class of intrinsics is designed to be generic and has
5591no specific purpose. </p>
5592</div>
5593
5594<!-- _______________________________________________________________________ -->
5595<div class="doc_subsubsection">
5596 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
5597</div>
5598
5599<div class="doc_text">
5600
5601<h5>Syntax:</h5>
5602<pre>
Tanya Lattnerd2e84422007-06-18 23:42:37 +00005603 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
Tanya Lattner6d806e92007-06-15 20:50:54 +00005604</pre>
5605
5606<h5>Overview:</h5>
5607
5608<p>
5609The '<tt>llvm.var.annotation</tt>' intrinsic
5610</p>
5611
5612<h5>Arguments:</h5>
5613
5614<p>
Tanya Lattnerd2e84422007-06-18 23:42:37 +00005615The first argument is a pointer to a value, the second is a pointer to a
5616global string, the third is a pointer to a global string which is the source
5617file name, and the last argument is the line number.
Tanya Lattner6d806e92007-06-15 20:50:54 +00005618</p>
5619
5620<h5>Semantics:</h5>
5621
5622<p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00005623This intrinsic allows annotation of local variables with arbitrary strings.
Tanya Lattner6d806e92007-06-15 20:50:54 +00005624This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00005625annotations. These have no other defined use, they are ignored by code
5626generation and optimization.
5627</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +00005628</div>
5629
Tanya Lattnerb6367882007-09-21 22:59:12 +00005630<!-- _______________________________________________________________________ -->
5631<div class="doc_subsubsection">
Tanya Lattnere1a8da02007-09-21 23:57:59 +00005632 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb6367882007-09-21 22:59:12 +00005633</div>
5634
5635<div class="doc_text">
5636
5637<h5>Syntax:</h5>
Tanya Lattner39cfba62007-09-21 23:56:27 +00005638<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
5639any integer bit width.
5640</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +00005641<pre>
Tanya Lattnerd3989a82007-09-22 00:03:01 +00005642 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5643 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5644 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5645 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5646 declare i256 @llvm.annotation.i256(i256 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
Tanya Lattnerb6367882007-09-21 22:59:12 +00005647</pre>
5648
5649<h5>Overview:</h5>
Tanya Lattner39cfba62007-09-21 23:56:27 +00005650
5651<p>
5652The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb6367882007-09-21 22:59:12 +00005653</p>
5654
5655<h5>Arguments:</h5>
5656
5657<p>
5658The first argument is an integer value (result of some expression),
5659the second is a pointer to a global string, the third is a pointer to a global
5660string which is the source file name, and the last argument is the line number.
Tanya Lattner39cfba62007-09-21 23:56:27 +00005661It returns the value of the first argument.
Tanya Lattnerb6367882007-09-21 22:59:12 +00005662</p>
5663
5664<h5>Semantics:</h5>
5665
5666<p>
5667This intrinsic allows annotations to be put on arbitrary expressions
5668with arbitrary strings. This can be useful for special purpose optimizations
5669that want to look for these annotations. These have no other defined use, they
5670are ignored by code generation and optimization.
5671</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00005672
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00005673<!-- _______________________________________________________________________ -->
5674<div class="doc_subsubsection">
5675 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
5676</div>
5677
5678<div class="doc_text">
5679
5680<h5>Syntax:</h5>
5681<pre>
5682 declare void @llvm.trap()
5683</pre>
5684
5685<h5>Overview:</h5>
5686
5687<p>
5688The '<tt>llvm.trap</tt>' intrinsic
5689</p>
5690
5691<h5>Arguments:</h5>
5692
5693<p>
5694None
5695</p>
5696
5697<h5>Semantics:</h5>
5698
5699<p>
5700This intrinsics is lowered to the target dependent trap instruction. If the
5701target does not have a trap instruction, this intrinsic will be lowered to the
5702call of the abort() function.
5703</p>
5704</div>
5705
Chris Lattner00950542001-06-06 20:29:01 +00005706<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00005707<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00005708<address>
5709 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
5710 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
5711 <a href="http://validator.w3.org/check/referer"><img
Chris Lattnerc7d3ab32008-01-04 04:33:49 +00005712 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00005713
5714 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +00005715 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00005716 Last modified: $Date$
5717</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +00005718
Misha Brukman9d0919f2003-11-08 01:05:38 +00005719</body>
5720</html>