blob: 74678c7be8e2ee01406098a49ce43130ff184d7b [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">
Eric Christopher6c7e8a02009-12-05 02:46:03 +00008 <meta name="description"
Reid Spencer3921c742004-08-26 20:44:00 +00009 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
NAKAMURA Takumi05d02652011-04-18 23:59:50 +000015<h1>LLVM Language Reference Manual</h1>
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>
Bill Wendling3d10a5a2009-07-20 01:03:30 +000023 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling987e7eb2009-07-20 02:41:50 +000025 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
Bill Wendling5e721d72010-07-01 21:55:59 +000027 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling55ae5152010-08-20 22:05:50 +000028 <li><a href="#linkage_linker_private_weak_def_auto">'<tt>linker_private_weak_def_auto</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +000029 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
30 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
31 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
32 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
33 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
34 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
35 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner5a2d8752009-10-10 18:26:06 +000036 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +000037 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
38 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +000041 </ol>
42 </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +000043 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnere7886e42009-01-11 20:53:49 +000044 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000045 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +000046 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +000047 <li><a href="#aliasstructure">Aliases</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +000048 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerca86e162006-12-31 07:07:53 +000049 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel2c9c3e72008-09-26 23:51:19 +000050 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +000051 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +000052 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencerde151942007-02-19 23:54:10 +000053 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman556ca272009-07-27 18:07:55 +000054 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +000055 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000056 </ol>
57 </li>
Chris Lattner00950542001-06-06 20:29:01 +000058 <li><a href="#typesystem">Type System</a>
59 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +000060 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +000061 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +000062 <ol>
Nick Lewyckyec38da42009-09-27 00:45:11 +000063 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +000064 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen21fe99b2010-10-01 00:48:59 +000065 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +000066 <li><a href="#t_void">Void Type</a></li>
67 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky7a0370f2009-05-30 05:06:04 +000068 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000069 </ol>
70 </li>
Chris Lattner00950542001-06-06 20:29:01 +000071 <li><a href="#t_derived">Derived Types</a>
72 <ol>
Chris Lattnerfdfeb692010-02-12 20:49:41 +000073 <li><a href="#t_aggregate">Aggregate Types</a>
74 <ol>
75 <li><a href="#t_array">Array Type</a></li>
76 <li><a href="#t_struct">Structure Type</a></li>
77 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Chris Lattnerfdfeb692010-02-12 20:49:41 +000078 <li><a href="#t_vector">Vector Type</a></li>
79 </ol>
80 </li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000081 <li><a href="#t_function">Function Type</a></li>
82 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner69c11bb2005-04-25 17:34:15 +000083 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000084 </ol>
85 </li>
Chris Lattner242d61d2009-02-02 07:32:36 +000086 <li><a href="#t_uprefs">Type Up-references</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000087 </ol>
88 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +000089 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +000090 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +000091 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner70882792009-02-28 18:32:25 +000092 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +000093 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
94 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanfff6c532010-04-22 23:14:21 +000095 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattnerf9d078e2009-10-27 21:19:13 +000096 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +000097 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattnerc3f59762004-12-09 17:30:23 +000098 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +000099 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000104 </ol>
105 </li>
Chris Lattner857755c2009-07-20 05:55:19 +0000106 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
107 <ol>
108 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner401e10c2009-07-20 06:14:25 +0000109 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
110 Global Variable</a></li>
Chris Lattner857755c2009-07-20 05:55:19 +0000111 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
112 Global Variable</a></li>
113 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
114 Global Variable</a></li>
115 </ol>
116 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000117 <li><a href="#instref">Instruction Reference</a>
118 <ol>
119 <li><a href="#terminators">Terminator Instructions</a>
120 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000121 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
122 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000123 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerab21db72009-10-28 00:19:10 +0000124 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000125 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000126 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000127 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000128 </ol>
129 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000130 <li><a href="#binaryops">Binary Operations</a>
131 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000132 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000133 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000134 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000135 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000136 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000137 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000138 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
139 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
140 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000141 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
142 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
143 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000144 </ol>
145 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000146 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
147 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000148 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
149 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
150 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000151 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000152 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000153 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000154 </ol>
155 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000156 <li><a href="#vectorops">Vector Operations</a>
157 <ol>
158 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
159 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
160 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000161 </ol>
162 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000163 <li><a href="#aggregateops">Aggregate Operations</a>
164 <ol>
165 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
166 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
167 </ol>
168 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000169 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000170 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000171 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000172 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
173 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
174 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000175 </ol>
176 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000177 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000178 <ol>
179 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
180 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
181 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
182 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
183 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000184 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
185 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
186 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
187 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000188 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
189 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000190 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000191 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000192 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000193 <li><a href="#otherops">Other Operations</a>
194 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000195 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
196 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000197 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000198 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000199 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000200 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000201 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000202 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000203 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000204 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000205 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000206 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000207 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
208 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000209 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
210 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
211 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000212 </ol>
213 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000214 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
215 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000216 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
217 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
218 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000219 </ol>
220 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000221 <li><a href="#int_codegen">Code Generator Intrinsics</a>
222 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000223 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
224 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
225 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
226 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
227 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
228 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000229 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000230 </ol>
231 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000232 <li><a href="#int_libc">Standard C Library Intrinsics</a>
233 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000234 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
235 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
238 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000239 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohman08b280b2011-05-27 00:36:31 +0000242 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
243 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000244 </ol>
245 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000246 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000247 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000248 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000249 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
250 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
251 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000252 </ol>
253 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000254 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
255 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000256 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
257 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
258 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
259 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
260 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendling41b485c2009-02-08 23:00:09 +0000261 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000262 </ol>
263 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000264 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
265 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000266 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
267 <li><a href="#int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a></li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000268 </ol>
269 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000270 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000271 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000272 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000273 <ol>
274 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000275 </ol>
276 </li>
Bill Wendling3c44f5b2008-11-18 22:10:53 +0000277 <li><a href="#int_atomics">Atomic intrinsics</a>
278 <ol>
279 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
280 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
281 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
282 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
283 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
284 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
285 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
286 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
287 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
288 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
289 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
290 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
291 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
292 </ol>
293 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000294 <li><a href="#int_memorymarkers">Memory Use Markers</a>
295 <ol>
296 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
297 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
298 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
299 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
300 </ol>
301 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000302 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000303 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000304 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000305 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000306 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000307 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000308 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000309 '<tt>llvm.trap</tt>' Intrinsic</a></li>
310 <li><a href="#int_stackprotector">
311 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher0e671492009-11-30 08:03:53 +0000312 <li><a href="#int_objectsize">
313 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000314 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000315 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000316 </ol>
317 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000318</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000319
320<div class="doc_author">
321 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
322 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000323</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000324
Chris Lattner00950542001-06-06 20:29:01 +0000325<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000326<h2><a name="abstract">Abstract</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000327<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000328
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000329<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000330
331<p>This document is a reference manual for the LLVM assembly language. LLVM is
332 a Static Single Assignment (SSA) based representation that provides type
333 safety, low-level operations, flexibility, and the capability of representing
334 'all' high-level languages cleanly. It is the common code representation
335 used throughout all phases of the LLVM compilation strategy.</p>
336
Misha Brukman9d0919f2003-11-08 01:05:38 +0000337</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000338
Chris Lattner00950542001-06-06 20:29:01 +0000339<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000340<h2><a name="introduction">Introduction</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000341<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000342
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000343<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000344
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000345<p>The LLVM code representation is designed to be used in three different forms:
346 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
347 for fast loading by a Just-In-Time compiler), and as a human readable
348 assembly language representation. This allows LLVM to provide a powerful
349 intermediate representation for efficient compiler transformations and
350 analysis, while providing a natural means to debug and visualize the
351 transformations. The three different forms of LLVM are all equivalent. This
352 document describes the human readable representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000353
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000354<p>The LLVM representation aims to be light-weight and low-level while being
355 expressive, typed, and extensible at the same time. It aims to be a
356 "universal IR" of sorts, by being at a low enough level that high-level ideas
357 may be cleanly mapped to it (similar to how microprocessors are "universal
358 IR's", allowing many source languages to be mapped to them). By providing
359 type information, LLVM can be used as the target of optimizations: for
360 example, through pointer analysis, it can be proven that a C automatic
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000361 variable is never accessed outside of the current function, allowing it to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000362 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000363
Chris Lattner00950542001-06-06 20:29:01 +0000364<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000365<h4>
366 <a name="wellformed">Well-Formedness</a>
367</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000368
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000369<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000370
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000371<p>It is important to note that this document describes 'well formed' LLVM
372 assembly language. There is a difference between what the parser accepts and
373 what is considered 'well formed'. For example, the following instruction is
374 syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000375
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000376<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000377%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000378</pre>
379
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000380<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
381 LLVM infrastructure provides a verification pass that may be used to verify
382 that an LLVM module is well formed. This pass is automatically run by the
383 parser after parsing input assembly and by the optimizer before it outputs
384 bitcode. The violations pointed out by the verifier pass indicate bugs in
385 transformation passes or input to the parser.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000386
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000387</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000388
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000389</div>
390
Chris Lattnercc689392007-10-03 17:34:29 +0000391<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000392
Chris Lattner00950542001-06-06 20:29:01 +0000393<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000394<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner00950542001-06-06 20:29:01 +0000395<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000396
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000397<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000398
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000399<p>LLVM identifiers come in two basic types: global and local. Global
400 identifiers (functions, global variables) begin with the <tt>'@'</tt>
401 character. Local identifiers (register names, types) begin with
402 the <tt>'%'</tt> character. Additionally, there are three different formats
403 for identifiers, for different purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000404
Chris Lattner00950542001-06-06 20:29:01 +0000405<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000406 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000407 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
408 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
409 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
410 other characters in their names can be surrounded with quotes. Special
411 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
412 ASCII code for the character in hexadecimal. In this way, any character
413 can be used in a name value, even quotes themselves.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000414
Reid Spencer2c452282007-08-07 14:34:28 +0000415 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000416 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000417
Reid Spencercc16dc32004-12-09 18:02:53 +0000418 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000419 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000420</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000421
Reid Spencer2c452282007-08-07 14:34:28 +0000422<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000423 don't need to worry about name clashes with reserved words, and the set of
424 reserved words may be expanded in the future without penalty. Additionally,
425 unnamed identifiers allow a compiler to quickly come up with a temporary
426 variable without having to avoid symbol table conflicts.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000427
Chris Lattner261efe92003-11-25 01:02:51 +0000428<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000429 languages. There are keywords for different opcodes
430 ('<tt><a href="#i_add">add</a></tt>',
431 '<tt><a href="#i_bitcast">bitcast</a></tt>',
432 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
433 ('<tt><a href="#t_void">void</a></tt>',
434 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
435 reserved words cannot conflict with variable names, because none of them
436 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000437
438<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000439 '<tt>%X</tt>' by 8:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000440
Misha Brukman9d0919f2003-11-08 01:05:38 +0000441<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000442
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000443<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000444%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000445</pre>
446
Misha Brukman9d0919f2003-11-08 01:05:38 +0000447<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000448
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000449<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000450%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000451</pre>
452
Misha Brukman9d0919f2003-11-08 01:05:38 +0000453<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000454
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000455<pre class="doc_code">
Gabor Greifec58f752009-10-28 13:05:07 +0000456%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
457%1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000458%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000459</pre>
460
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000461<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
462 lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000463
Chris Lattner00950542001-06-06 20:29:01 +0000464<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000465 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000466 line.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000467
468 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000469 assigned to a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000470
Misha Brukman9d0919f2003-11-08 01:05:38 +0000471 <li>Unnamed temporaries are numbered sequentially</li>
472</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000473
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000474<p>It also shows a convention that we follow in this document. When
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000475 demonstrating instructions, we will follow an instruction with a comment that
476 defines the type and name of value produced. Comments are shown in italic
477 text.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000478
Misha Brukman9d0919f2003-11-08 01:05:38 +0000479</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000480
481<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000482<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattnerfa730212004-12-09 16:11:40 +0000483<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000484<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000485<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000486<h3>
487 <a name="modulestructure">Module Structure</a>
488</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000489
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000490<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000491
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000492<p>LLVM programs are composed of "Module"s, each of which is a translation unit
493 of the input programs. Each module consists of functions, global variables,
494 and symbol table entries. Modules may be combined together with the LLVM
495 linker, which merges function (and global variable) definitions, resolves
496 forward declarations, and merges symbol table entries. Here is an example of
497 the "hello world" module:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000498
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000499<pre class="doc_code">
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000500<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckydb9cd762011-01-29 01:09:53 +0000501<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a>&nbsp;<a href="#globalvars">constant</a>&nbsp;<a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000502
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000503<i>; External declaration of the puts function</i>&nbsp;
504<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000505
506<i>; Definition of main function</i>
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000507define i32 @main() { <i>; i32()* </i>&nbsp;
508 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
509 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8*</i>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000510
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000511 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
512 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
513 <a href="#i_ret">ret</a> i32 0&nbsp;
514}
Devang Patelcd1fd252010-01-11 19:35:55 +0000515
516<i>; Named metadata</i>
517!1 = metadata !{i32 41}
518!foo = !{!1, null}
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000519</pre>
Chris Lattnerfa730212004-12-09 16:11:40 +0000520
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000521<p>This example is made up of a <a href="#globalvars">global variable</a> named
Devang Patelcd1fd252010-01-11 19:35:55 +0000522 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000523 a <a href="#functionstructure">function definition</a> for
Devang Patelcd1fd252010-01-11 19:35:55 +0000524 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
525 "<tt>foo"</tt>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000526
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000527<p>In general, a module is made up of a list of global values, where both
528 functions and global variables are global values. Global values are
529 represented by a pointer to a memory location (in this case, a pointer to an
530 array of char, and a pointer to a function), and have one of the
531 following <a href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000532
Chris Lattnere5d947b2004-12-09 16:36:40 +0000533</div>
534
535<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000536<h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000537 <a name="linkage">Linkage Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000538</h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000539
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000540<div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000541
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000542<p>All Global Variables and Functions have one of the following types of
543 linkage:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000544
545<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000546 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000547 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
548 by objects in the current module. In particular, linking code into a
549 module with an private global value may cause the private to be renamed as
550 necessary to avoid collisions. Because the symbol is private to the
551 module, all references can be updated. This doesn't show up in any symbol
552 table in the object file.</dd>
Rafael Espindolabb46f522009-01-15 20:18:42 +0000553
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000554 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000555 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
556 assembler and evaluated by the linker. Unlike normal strong symbols, they
557 are removed by the linker from the final linked image (executable or
558 dynamic library).</dd>
559
560 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
561 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
562 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
563 linker. The symbols are removed by the linker from the final linked image
564 (executable or dynamic library).</dd>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000565
Bill Wendling55ae5152010-08-20 22:05:50 +0000566 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
567 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
568 of the object is not taken. For instance, functions that had an inline
569 definition, but the compiler decided not to inline it. Note,
570 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
571 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
572 visibility. The symbols are removed by the linker from the final linked
573 image (executable or dynamic library).</dd>
574
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000575 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling07d31772010-06-29 22:34:52 +0000576 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000577 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
578 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000579
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000580 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000581 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000582 into the object file corresponding to the LLVM module. They exist to
583 allow inlining and other optimizations to take place given knowledge of
584 the definition of the global, which is known to be somewhere outside the
585 module. Globals with <tt>available_externally</tt> linkage are allowed to
586 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
587 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000588
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000589 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattner4887bd82007-01-14 06:51:48 +0000590 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner873187c2010-01-09 19:15:14 +0000591 the same name when linkage occurs. This can be used to implement
592 some forms of inline functions, templates, or other code which must be
593 generated in each translation unit that uses it, but where the body may
594 be overridden with a more definitive definition later. Unreferenced
595 <tt>linkonce</tt> globals are allowed to be discarded. Note that
596 <tt>linkonce</tt> linkage does not actually allow the optimizer to
597 inline the body of this function into callers because it doesn't know if
598 this definition of the function is the definitive definition within the
599 program or whether it will be overridden by a stronger definition.
600 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
601 linkage.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000602
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000603 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000604 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
605 <tt>linkonce</tt> linkage, except that unreferenced globals with
606 <tt>weak</tt> linkage may not be discarded. This is used for globals that
607 are declared "weak" in C source code.</dd>
608
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000609 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000610 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
611 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
612 global scope.
613 Symbols with "<tt>common</tt>" linkage are merged in the same way as
614 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000615 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000616 must have a zero initializer, and may not be marked '<a
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000617 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
618 have common linkage.</dd>
Chris Lattner26d054d2009-08-05 05:21:07 +0000619
Chris Lattnere5d947b2004-12-09 16:36:40 +0000620
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000621 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000622 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000623 pointer to array type. When two global variables with appending linkage
624 are linked together, the two global arrays are appended together. This is
625 the LLVM, typesafe, equivalent of having the system linker append together
626 "sections" with identical names when .o files are linked.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000627
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000628 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000629 <dd>The semantics of this linkage follow the ELF object file model: the symbol
630 is weak until linked, if not linked, the symbol becomes null instead of
631 being an undefined reference.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000632
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000633 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
634 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000635 <dd>Some languages allow differing globals to be merged, such as two functions
636 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling5e721d72010-07-01 21:55:59 +0000637 that only equivalent globals are ever merged (the "one definition rule"
638 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000639 and <tt>weak_odr</tt> linkage types to indicate that the global will only
640 be merged with equivalent globals. These linkage types are otherwise the
641 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands667d4b82009-03-07 15:45:40 +0000642
Chris Lattnerfa730212004-12-09 16:11:40 +0000643 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000644 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000645 visible, meaning that it participates in linkage and can be used to
646 resolve external symbol references.</dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000647</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000648
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000649<p>The next two types of linkage are targeted for Microsoft Windows platform
650 only. They are designed to support importing (exporting) symbols from (to)
651 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000652
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000653<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000654 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000655 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000656 or variable via a global pointer to a pointer that is set up by the DLL
657 exporting the symbol. On Microsoft Windows targets, the pointer name is
658 formed by combining <code>__imp_</code> and the function or variable
659 name.</dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000660
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000661 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000662 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000663 pointer to a pointer in a DLL, so that it can be referenced with the
664 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
665 name is formed by combining <code>__imp_</code> and the function or
666 variable name.</dd>
Chris Lattnerfa730212004-12-09 16:11:40 +0000667</dl>
668
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000669<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
670 another module defined a "<tt>.LC0</tt>" variable and was linked with this
671 one, one of the two would be renamed, preventing a collision. Since
672 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
673 declarations), they are accessible outside of the current module.</p>
674
675<p>It is illegal for a function <i>declaration</i> to have any linkage type
676 other than "externally visible", <tt>dllimport</tt>
677 or <tt>extern_weak</tt>.</p>
678
Duncan Sands667d4b82009-03-07 15:45:40 +0000679<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000680 or <tt>weak_odr</tt> linkages.</p>
681
Chris Lattnerfa730212004-12-09 16:11:40 +0000682</div>
683
684<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000685<h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000686 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000687</h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000688
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000689<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000690
691<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000692 and <a href="#i_invoke">invokes</a> can all have an optional calling
693 convention specified for the call. The calling convention of any pair of
694 dynamic caller/callee must match, or the behavior of the program is
695 undefined. The following calling conventions are supported by LLVM, and more
696 may be added in the future:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000697
698<dl>
699 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000700 <dd>This calling convention (the default if no other calling convention is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000701 specified) matches the target C calling conventions. This calling
702 convention supports varargs function calls and tolerates some mismatch in
703 the declared prototype and implemented declaration of the function (as
704 does normal C).</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000705
706 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000707 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000708 (e.g. by passing things in registers). This calling convention allows the
709 target to use whatever tricks it wants to produce fast code for the
710 target, without having to conform to an externally specified ABI
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000711 (Application Binary Interface).
712 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattner29689432010-03-11 00:22:57 +0000713 when this or the GHC convention is used.</a> This calling convention
714 does not support varargs and requires the prototype of all callees to
715 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000716
717 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000718 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000719 as possible under the assumption that the call is not commonly executed.
720 As such, these calls often preserve all registers so that the call does
721 not break any live ranges in the caller side. This calling convention
722 does not support varargs and requires the prototype of all callees to
723 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000724
Chris Lattner29689432010-03-11 00:22:57 +0000725 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
726 <dd>This calling convention has been implemented specifically for use by the
727 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
728 It passes everything in registers, going to extremes to achieve this by
729 disabling callee save registers. This calling convention should not be
730 used lightly but only for specific situations such as an alternative to
731 the <em>register pinning</em> performance technique often used when
732 implementing functional programming languages.At the moment only X86
733 supports this convention and it has the following limitations:
734 <ul>
735 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
736 floating point types are supported.</li>
737 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
738 6 floating point parameters.</li>
739 </ul>
740 This calling convention supports
741 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
742 requires both the caller and callee are using it.
743 </dd>
744
Chris Lattnercfe6b372005-05-07 01:46:40 +0000745 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000746 <dd>Any calling convention may be specified by number, allowing
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000747 target-specific calling conventions to be used. Target specific calling
748 conventions start at 64.</dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000749</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000750
751<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000752 support Pascal conventions or any other well-known target-independent
753 convention.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000754
755</div>
756
757<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000758<h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000759 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000760</h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000761
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000762<div>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000763
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000764<p>All Global Variables and Functions have one of the following visibility
765 styles:</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000766
767<dl>
768 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000769 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000770 that the declaration is visible to other modules and, in shared libraries,
771 means that the declared entity may be overridden. On Darwin, default
772 visibility means that the declaration is visible to other modules. Default
773 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000774
775 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000776 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000777 object if they are in the same shared object. Usually, hidden visibility
778 indicates that the symbol will not be placed into the dynamic symbol
779 table, so no other module (executable or shared library) can reference it
780 directly.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000781
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000782 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000783 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000784 the dynamic symbol table, but that references within the defining module
785 will bind to the local symbol. That is, the symbol cannot be overridden by
786 another module.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000787</dl>
788
789</div>
790
791<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000792<h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000793 <a name="namedtypes">Named Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000794</h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000795
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000796<div>
Chris Lattnere7886e42009-01-11 20:53:49 +0000797
798<p>LLVM IR allows you to specify name aliases for certain types. This can make
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000799 it easier to read the IR and make the IR more condensed (particularly when
800 recursive types are involved). An example of a name specification is:</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000801
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000802<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000803%mytype = type { %mytype*, i32 }
804</pre>
Chris Lattnere7886e42009-01-11 20:53:49 +0000805
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000806<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattnerdc65f222010-08-17 23:26:04 +0000807 "<a href="#t_void">void</a>". Type name aliases may be used anywhere a type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000808 is expected with the syntax "%mytype".</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000809
810<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000811 and that you can therefore specify multiple names for the same type. This
812 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
813 uses structural typing, the name is not part of the type. When printing out
814 LLVM IR, the printer will pick <em>one name</em> to render all types of a
815 particular shape. This means that if you have code where two different
816 source types end up having the same LLVM type, that the dumper will sometimes
817 print the "wrong" or unexpected type. This is an important design point and
818 isn't going to change.</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000819
820</div>
821
Chris Lattnere7886e42009-01-11 20:53:49 +0000822<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000823<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000824 <a name="globalvars">Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000825</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000826
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000827<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000828
Chris Lattner3689a342005-02-12 19:30:21 +0000829<p>Global variables define regions of memory allocated at compilation time
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000830 instead of run-time. Global variables may optionally be initialized, may
831 have an explicit section to be placed in, and may have an optional explicit
832 alignment specified. A variable may be defined as "thread_local", which
833 means that it will not be shared by threads (each thread will have a
834 separated copy of the variable). A variable may be defined as a global
835 "constant," which indicates that the contents of the variable
836 will <b>never</b> be modified (enabling better optimization, allowing the
837 global data to be placed in the read-only section of an executable, etc).
838 Note that variables that need runtime initialization cannot be marked
839 "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000840
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000841<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
842 constant, even if the final definition of the global is not. This capability
843 can be used to enable slightly better optimization of the program, but
844 requires the language definition to guarantee that optimizations based on the
845 'constantness' are valid for the translation units that do not include the
846 definition.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000847
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000848<p>As SSA values, global variables define pointer values that are in scope
849 (i.e. they dominate) all basic blocks in the program. Global variables
850 always define a pointer to their "content" type because they describe a
851 region of memory, and all memory objects in LLVM are accessed through
852 pointers.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000853
Rafael Espindolabea46262011-01-08 16:42:36 +0000854<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
855 that the address is not significant, only the content. Constants marked
Rafael Espindolaa5eaa862011-01-15 08:20:57 +0000856 like this can be merged with other constants if they have the same
857 initializer. Note that a constant with significant address <em>can</em>
858 be merged with a <tt>unnamed_addr</tt> constant, the result being a
859 constant whose address is significant.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000860
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000861<p>A global variable may be declared to reside in a target-specific numbered
862 address space. For targets that support them, address spaces may affect how
863 optimizations are performed and/or what target instructions are used to
864 access the variable. The default address space is zero. The address space
865 qualifier must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000866
Chris Lattner88f6c462005-11-12 00:45:07 +0000867<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000868 supports it, it will emit globals to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000869
Chris Lattnerce99fa92010-04-28 00:13:42 +0000870<p>An explicit alignment may be specified for a global, which must be a power
871 of 2. If not present, or if the alignment is set to zero, the alignment of
872 the global is set by the target to whatever it feels convenient. If an
873 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner2d4b8ee2010-04-28 00:31:12 +0000874 alignment. Targets and optimizers are not allowed to over-align the global
875 if the global has an assigned section. In this case, the extra alignment
876 could be observable: for example, code could assume that the globals are
877 densely packed in their section and try to iterate over them as an array,
878 alignment padding would break this iteration.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000879
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000880<p>For example, the following defines a global in a numbered address space with
881 an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000882
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000883<pre class="doc_code">
Dan Gohman398873c2009-01-11 00:40:00 +0000884@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000885</pre>
886
Chris Lattnerfa730212004-12-09 16:11:40 +0000887</div>
888
889
890<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000891<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000892 <a name="functionstructure">Functions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000893</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000894
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000895<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000896
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000897<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000898 optional <a href="#linkage">linkage type</a>, an optional
899 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000900 <a href="#callingconv">calling convention</a>,
901 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000902 <a href="#paramattrs">parameter attribute</a> for the return type, a function
903 name, a (possibly empty) argument list (each with optional
904 <a href="#paramattrs">parameter attributes</a>), optional
905 <a href="#fnattrs">function attributes</a>, an optional section, an optional
906 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
907 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000908
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000909<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
910 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000911 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000912 <a href="#callingconv">calling convention</a>,
913 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000914 <a href="#paramattrs">parameter attribute</a> for the return type, a function
915 name, a possibly empty list of arguments, an optional alignment, and an
916 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000917
Chris Lattnerd3eda892008-08-05 18:29:16 +0000918<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000919 (Control Flow Graph) for the function. Each basic block may optionally start
920 with a label (giving the basic block a symbol table entry), contains a list
921 of instructions, and ends with a <a href="#terminators">terminator</a>
922 instruction (such as a branch or function return).</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000923
Chris Lattner4a3c9012007-06-08 16:52:14 +0000924<p>The first basic block in a function is special in two ways: it is immediately
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000925 executed on entrance to the function, and it is not allowed to have
926 predecessor basic blocks (i.e. there can not be any branches to the entry
927 block of a function). Because the block can have no predecessors, it also
928 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000929
Chris Lattner88f6c462005-11-12 00:45:07 +0000930<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000931 supports it, it will emit functions to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000932
Chris Lattner2cbdc452005-11-06 08:02:57 +0000933<p>An explicit alignment may be specified for a function. If not present, or if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000934 the alignment is set to zero, the alignment of the function is set by the
935 target to whatever it feels convenient. If an explicit alignment is
936 specified, the function is forced to have at least that much alignment. All
937 alignments must be a power of 2.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000938
Rafael Espindolabea46262011-01-08 16:42:36 +0000939<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
940 be significant and two identical functions can be merged</p>.
941
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000942<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000943<pre class="doc_code">
Chris Lattner50ad45c2008-10-13 16:55:18 +0000944define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000945 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
946 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
947 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
948 [<a href="#gc">gc</a>] { ... }
949</pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000950
Chris Lattnerfa730212004-12-09 16:11:40 +0000951</div>
952
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000953<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000954<h3>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000955 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000956</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000957
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000958<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000959
960<p>Aliases act as "second name" for the aliasee value (which can be either
961 function, global variable, another alias or bitcast of global value). Aliases
962 may have an optional <a href="#linkage">linkage type</a>, and an
963 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000964
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000965<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000966<pre class="doc_code">
Duncan Sands0b23ac12008-09-12 20:48:21 +0000967@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000968</pre>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000969
970</div>
971
Chris Lattner4e9aba72006-01-23 23:23:47 +0000972<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000973<h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000974 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000975</h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000976
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000977<div>
Devang Patelcd1fd252010-01-11 19:35:55 +0000978
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000979<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman872814a2010-07-21 18:54:18 +0000980 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000981 a named metadata.</p>
Devang Patelcd1fd252010-01-11 19:35:55 +0000982
983<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000984<pre class="doc_code">
Dan Gohman872814a2010-07-21 18:54:18 +0000985; Some unnamed metadata nodes, which are referenced by the named metadata.
986!0 = metadata !{metadata !"zero"}
Devang Patelcd1fd252010-01-11 19:35:55 +0000987!1 = metadata !{metadata !"one"}
Dan Gohman872814a2010-07-21 18:54:18 +0000988!2 = metadata !{metadata !"two"}
Dan Gohman1005bc52010-07-13 19:48:13 +0000989; A named metadata.
Dan Gohman872814a2010-07-21 18:54:18 +0000990!name = !{!0, !1, !2}
Devang Patelcd1fd252010-01-11 19:35:55 +0000991</pre>
Devang Patelcd1fd252010-01-11 19:35:55 +0000992
993</div>
994
995<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000996<h3>
997 <a name="paramattrs">Parameter Attributes</a>
998</h3>
Reid Spencerca86e162006-12-31 07:07:53 +0000999
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001000<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001001
1002<p>The return type and each parameter of a function type may have a set of
1003 <i>parameter attributes</i> associated with them. Parameter attributes are
1004 used to communicate additional information about the result or parameters of
1005 a function. Parameter attributes are considered to be part of the function,
1006 not of the function type, so functions with different parameter attributes
1007 can have the same function type.</p>
1008
1009<p>Parameter attributes are simple keywords that follow the type specified. If
1010 multiple parameter attributes are needed, they are space separated. For
1011 example:</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001012
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00001013<pre class="doc_code">
Nick Lewyckyb6a7d252009-02-15 23:06:14 +00001014declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner66d922c2008-10-04 18:33:34 +00001015declare i32 @atoi(i8 zeroext)
1016declare signext i8 @returns_signed_char()
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001017</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001018
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001019<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1020 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +00001021
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001022<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner47507de2008-01-11 06:20:47 +00001023
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001024<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001025 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001026 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichebe81732011-03-16 22:20:18 +00001027 should be zero-extended to the extent required by the target's ABI (which
1028 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1029 parameter) or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +00001030
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001031 <dt><tt><b>signext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001032 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich9e69ff92011-03-17 14:21:58 +00001033 should be sign-extended to the extent required by the target's ABI (which
1034 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1035 return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +00001036
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001037 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001038 <dd>This indicates that this parameter or return value should be treated in a
1039 special target-dependent fashion during while emitting code for a function
1040 call or return (usually, by putting it in a register as opposed to memory,
1041 though some targets use it to distinguish between two different kinds of
1042 registers). Use of this attribute is target-specific.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +00001043
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001044 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnera6fd81d2010-11-20 23:49:06 +00001045 <dd><p>This indicates that the pointer parameter should really be passed by
1046 value to the function. The attribute implies that a hidden copy of the
1047 pointee
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001048 is made between the caller and the callee, so the callee is unable to
1049 modify the value in the callee. This attribute is only valid on LLVM
1050 pointer arguments. It is generally used to pass structs and arrays by
1051 value, but is also valid on pointers to scalars. The copy is considered
1052 to belong to the caller not the callee (for example,
1053 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1054 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnera6fd81d2010-11-20 23:49:06 +00001055 values.</p>
1056
1057 <p>The byval attribute also supports specifying an alignment with
1058 the align attribute. It indicates the alignment of the stack slot to
1059 form and the known alignment of the pointer specified to the call site. If
1060 the alignment is not specified, then the code generator makes a
1061 target-specific assumption.</p></dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001062
Dan Gohmanff235352010-07-02 23:18:08 +00001063 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001064 <dd>This indicates that the pointer parameter specifies the address of a
1065 structure that is the return value of the function in the source program.
1066 This pointer must be guaranteed by the caller to be valid: loads and
1067 stores to the structure may be assumed by the callee to not to trap. This
1068 may only be applied to the first parameter. This is not a valid attribute
1069 for return values. </dd>
1070
Dan Gohmanff235352010-07-02 23:18:08 +00001071 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohman1e109622010-07-02 18:41:32 +00001072 <dd>This indicates that pointer values
1073 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmanefca7f92010-07-02 23:46:54 +00001074 value do not alias pointer values which are not <i>based</i> on it,
1075 ignoring certain "irrelevant" dependencies.
1076 For a call to the parent function, dependencies between memory
1077 references from before or after the call and from those during the call
1078 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1079 return value used in that call.
Dan Gohman1e109622010-07-02 18:41:32 +00001080 The caller shares the responsibility with the callee for ensuring that
1081 these requirements are met.
1082 For further details, please see the discussion of the NoAlias response in
Dan Gohmanff70fe42010-07-06 15:26:33 +00001083 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1084<br>
John McCall191d4ee2010-07-06 21:07:14 +00001085 Note that this definition of <tt>noalias</tt> is intentionally
1086 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner211244a2010-07-06 20:51:35 +00001087 arguments, though it is slightly weaker.
Dan Gohmanff70fe42010-07-06 15:26:33 +00001088<br>
1089 For function return values, C99's <tt>restrict</tt> is not meaningful,
1090 while LLVM's <tt>noalias</tt> is.
1091 </dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001092
Dan Gohmanff235352010-07-02 23:18:08 +00001093 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001094 <dd>This indicates that the callee does not make any copies of the pointer
1095 that outlive the callee itself. This is not a valid attribute for return
1096 values.</dd>
1097
Dan Gohmanff235352010-07-02 23:18:08 +00001098 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001099 <dd>This indicates that the pointer parameter can be excised using the
1100 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1101 attribute for return values.</dd>
1102</dl>
Reid Spencerca86e162006-12-31 07:07:53 +00001103
Reid Spencerca86e162006-12-31 07:07:53 +00001104</div>
1105
1106<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001107<h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +00001108 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001109</h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +00001110
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001111<div>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +00001112
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001113<p>Each function may specify a garbage collector name, which is simply a
1114 string:</p>
1115
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00001116<pre class="doc_code">
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001117define void @f() gc "name" { ... }
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001118</pre>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +00001119
1120<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001121 collector which will cause the compiler to alter its output in order to
1122 support the named garbage collection algorithm.</p>
1123
Gordon Henriksen80a75bf2007-12-10 03:18:06 +00001124</div>
1125
1126<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001127<h3>
Devang Patel2c9c3e72008-09-26 23:51:19 +00001128 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001129</h3>
Devang Patelf8b94812008-09-04 23:05:13 +00001130
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001131<div>
Devang Patel2c9c3e72008-09-26 23:51:19 +00001132
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001133<p>Function attributes are set to communicate additional information about a
1134 function. Function attributes are considered to be part of the function, not
1135 of the function type, so functions with different parameter attributes can
1136 have the same function type.</p>
Devang Patel2c9c3e72008-09-26 23:51:19 +00001137
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001138<p>Function attributes are simple keywords that follow the type specified. If
1139 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelf8b94812008-09-04 23:05:13 +00001140
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00001141<pre class="doc_code">
Devang Patel2c9c3e72008-09-26 23:51:19 +00001142define void @f() noinline { ... }
1143define void @f() alwaysinline { ... }
1144define void @f() alwaysinline optsize { ... }
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001145define void @f() optsize { ... }
Bill Wendlinge36dccc2008-09-07 10:26:33 +00001146</pre>
Devang Patelf8b94812008-09-04 23:05:13 +00001147
Bill Wendlinge36dccc2008-09-07 10:26:33 +00001148<dl>
Charles Davis1e063d12010-02-12 00:31:15 +00001149 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1150 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1151 the backend should forcibly align the stack pointer. Specify the
1152 desired alignment, which must be a power of two, in parentheses.
1153
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001154 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001155 <dd>This attribute indicates that the inliner should attempt to inline this
1156 function into callers whenever possible, ignoring any active inlining size
1157 threshold for this caller.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +00001158
Charles Davis970bfcc2010-10-25 15:37:09 +00001159 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis6f12e292010-10-25 16:29:03 +00001160 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis0076d202010-10-25 19:07:39 +00001161 meaning the function can be patched and/or hooked even while it is
1162 loaded into memory. On x86, the function prologue will be preceded
1163 by six bytes of padding and will begin with a two-byte instruction.
1164 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1165 higher were compiled in this fashion.</dd>
Charles Davis970bfcc2010-10-25 15:37:09 +00001166
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +00001167 <dt><tt><b>inlinehint</b></tt></dt>
1168 <dd>This attribute indicates that the source code contained a hint that inlining
1169 this function is desirable (such as the "inline" keyword in C/C++). It
1170 is just a hint; it imposes no requirements on the inliner.</dd>
1171
Nick Lewycky76ec37a2010-07-06 18:24:09 +00001172 <dt><tt><b>naked</b></tt></dt>
1173 <dd>This attribute disables prologue / epilogue emission for the function.
1174 This can have very system-specific consequences.</dd>
1175
1176 <dt><tt><b>noimplicitfloat</b></tt></dt>
1177 <dd>This attributes disables implicit floating point instructions.</dd>
1178
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001179 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001180 <dd>This attribute indicates that the inliner should never inline this
1181 function in any situation. This attribute may not be used together with
1182 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +00001183
Nick Lewycky76ec37a2010-07-06 18:24:09 +00001184 <dt><tt><b>noredzone</b></tt></dt>
1185 <dd>This attribute indicates that the code generator should not use a red
1186 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +00001187
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001188 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001189 <dd>This function attribute indicates that the function never returns
1190 normally. This produces undefined behavior at runtime if the function
1191 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +00001192
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001193 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001194 <dd>This function attribute indicates that the function never returns with an
1195 unwind or exceptional control flow. If the function does unwind, its
1196 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +00001197
Nick Lewycky76ec37a2010-07-06 18:24:09 +00001198 <dt><tt><b>optsize</b></tt></dt>
1199 <dd>This attribute suggests that optimization passes and code generator passes
1200 make choices that keep the code size of this function low, and otherwise
1201 do optimizations specifically to reduce code size.</dd>
1202
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001203 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001204 <dd>This attribute indicates that the function computes its result (or decides
1205 to unwind an exception) based strictly on its arguments, without
1206 dereferencing any pointer arguments or otherwise accessing any mutable
1207 state (e.g. memory, control registers, etc) visible to caller functions.
1208 It does not write through any pointer arguments
1209 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1210 changes any state visible to callers. This means that it cannot unwind
1211 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1212 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +00001213
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001214 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001215 <dd>This attribute indicates that the function does not write through any
1216 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1217 arguments) or otherwise modify any state (e.g. memory, control registers,
1218 etc) visible to caller functions. It may dereference pointer arguments
1219 and read state that may be set in the caller. A readonly function always
1220 returns the same value (or unwinds an exception identically) when called
1221 with the same set of arguments and global state. It cannot unwind an
1222 exception by calling the <tt>C++</tt> exception throwing methods, but may
1223 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +00001224
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001225 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001226 <dd>This attribute indicates that the function should emit a stack smashing
1227 protector. It is in the form of a "canary"&mdash;a random value placed on
1228 the stack before the local variables that's checked upon return from the
1229 function to see if it has been overwritten. A heuristic is used to
1230 determine if a function needs stack protectors or not.<br>
1231<br>
1232 If a function that has an <tt>ssp</tt> attribute is inlined into a
1233 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1234 function will have an <tt>ssp</tt> attribute.</dd>
1235
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001236 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001237 <dd>This attribute indicates that the function should <em>always</em> emit a
1238 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001239 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1240<br>
1241 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1242 function that doesn't have an <tt>sspreq</tt> attribute or which has
1243 an <tt>ssp</tt> attribute, then the resulting function will have
1244 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +00001245</dl>
1246
Devang Patelf8b94812008-09-04 23:05:13 +00001247</div>
1248
1249<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001250<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +00001251 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001252</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001253
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001254<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001255
1256<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1257 the GCC "file scope inline asm" blocks. These blocks are internally
1258 concatenated by LLVM and treated as a single unit, but may be separated in
1259 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001260
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00001261<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001262module asm "inline asm code goes here"
1263module asm "more can go here"
1264</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001265
1266<p>The strings can contain any character by escaping non-printable characters.
1267 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001268 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001269
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001270<p>The inline asm code is simply printed to the machine code .s file when
1271 assembly code is generated.</p>
1272
Chris Lattner4e9aba72006-01-23 23:23:47 +00001273</div>
Chris Lattnerfa730212004-12-09 16:11:40 +00001274
Reid Spencerde151942007-02-19 23:54:10 +00001275<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001276<h3>
Reid Spencerde151942007-02-19 23:54:10 +00001277 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001278</h3>
Reid Spencerde151942007-02-19 23:54:10 +00001279
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001280<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001281
Reid Spencerde151942007-02-19 23:54:10 +00001282<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001283 data is to be laid out in memory. The syntax for the data layout is
1284 simply:</p>
1285
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00001286<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001287target datalayout = "<i>layout specification</i>"
1288</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001289
1290<p>The <i>layout specification</i> consists of a list of specifications
1291 separated by the minus sign character ('-'). Each specification starts with
1292 a letter and may include other information after the letter to define some
1293 aspect of the data layout. The specifications accepted are as follows:</p>
1294
Reid Spencerde151942007-02-19 23:54:10 +00001295<dl>
1296 <dt><tt>E</tt></dt>
1297 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001298 bits with the most significance have the lowest address location.</dd>
1299
Reid Spencerde151942007-02-19 23:54:10 +00001300 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +00001301 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001302 the bits with the least significance have the lowest address
1303 location.</dd>
1304
Reid Spencerde151942007-02-19 23:54:10 +00001305 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001306 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001307 <i>preferred</i> alignments. All sizes are in bits. Specifying
1308 the <i>pref</i> alignment is optional. If omitted, the
1309 preceding <tt>:</tt> should be omitted too.</dd>
1310
Reid Spencerde151942007-02-19 23:54:10 +00001311 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1312 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001313 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1314
Reid Spencerde151942007-02-19 23:54:10 +00001315 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001316 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001317 <i>size</i>.</dd>
1318
Reid Spencerde151942007-02-19 23:54:10 +00001319 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001320 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +00001321 <i>size</i>. Only values of <i>size</i> that are supported by the target
1322 will work. 32 (float) and 64 (double) are supported on all targets;
1323 80 or 128 (different flavors of long double) are also supported on some
1324 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001325
Reid Spencerde151942007-02-19 23:54:10 +00001326 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1327 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001328 <i>size</i>.</dd>
1329
Daniel Dunbar87bde0b2009-06-08 22:17:53 +00001330 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1331 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001332 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +00001333
1334 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1335 <dd>This specifies a set of native integer widths for the target CPU
1336 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1337 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001338 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +00001339 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +00001340</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001341
Reid Spencerde151942007-02-19 23:54:10 +00001342<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +00001343 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001344 specifications in the <tt>datalayout</tt> keyword. The default specifications
1345 are given in this list:</p>
1346
Reid Spencerde151942007-02-19 23:54:10 +00001347<ul>
1348 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +00001349 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
Reid Spencerde151942007-02-19 23:54:10 +00001350 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1351 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1352 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1353 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +00001354 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencerde151942007-02-19 23:54:10 +00001355 alignment of 64-bits</li>
1356 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1357 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1358 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1359 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1360 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +00001361 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +00001362</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001363
1364<p>When LLVM is determining the alignment for a given type, it uses the
1365 following rules:</p>
1366
Reid Spencerde151942007-02-19 23:54:10 +00001367<ol>
1368 <li>If the type sought is an exact match for one of the specifications, that
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001369 specification is used.</li>
1370
Reid Spencerde151942007-02-19 23:54:10 +00001371 <li>If no match is found, and the type sought is an integer type, then the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001372 smallest integer type that is larger than the bitwidth of the sought type
1373 is used. If none of the specifications are larger than the bitwidth then
1374 the the largest integer type is used. For example, given the default
1375 specifications above, the i7 type will use the alignment of i8 (next
1376 largest) while both i65 and i256 will use the alignment of i64 (largest
1377 specified).</li>
1378
Reid Spencerde151942007-02-19 23:54:10 +00001379 <li>If no match is found, and the type sought is a vector type, then the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001380 largest vector type that is smaller than the sought vector type will be
1381 used as a fall back. This happens because &lt;128 x double&gt; can be
1382 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +00001383</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001384
Reid Spencerde151942007-02-19 23:54:10 +00001385</div>
Chris Lattnerfa730212004-12-09 16:11:40 +00001386
Dan Gohman556ca272009-07-27 18:07:55 +00001387<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001388<h3>
Dan Gohman556ca272009-07-27 18:07:55 +00001389 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001390</h3>
Dan Gohman556ca272009-07-27 18:07:55 +00001391
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001392<div>
Dan Gohman556ca272009-07-27 18:07:55 +00001393
Andreas Bolka55e459a2009-07-29 00:02:05 +00001394<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +00001395with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +00001396is undefined. Pointer values are associated with address ranges
1397according to the following rules:</p>
1398
1399<ul>
Dan Gohman1e109622010-07-02 18:41:32 +00001400 <li>A pointer value is associated with the addresses associated with
1401 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +00001402 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +00001403 range of the variable's storage.</li>
1404 <li>The result value of an allocation instruction is associated with
1405 the address range of the allocated storage.</li>
1406 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +00001407 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +00001408 <li>An integer constant other than zero or a pointer value returned
1409 from a function not defined within LLVM may be associated with address
1410 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +00001411 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +00001412 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +00001413</ul>
1414
1415<p>A pointer value is <i>based</i> on another pointer value according
1416 to the following rules:</p>
1417
1418<ul>
1419 <li>A pointer value formed from a
1420 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1421 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1422 <li>The result value of a
1423 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1424 of the <tt>bitcast</tt>.</li>
1425 <li>A pointer value formed by an
1426 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1427 pointer values that contribute (directly or indirectly) to the
1428 computation of the pointer's value.</li>
1429 <li>The "<i>based</i> on" relationship is transitive.</li>
1430</ul>
1431
1432<p>Note that this definition of <i>"based"</i> is intentionally
1433 similar to the definition of <i>"based"</i> in C99, though it is
1434 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +00001435
1436<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +00001437<tt><a href="#i_load">load</a></tt> merely indicates the size and
1438alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +00001439interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +00001440<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1441and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +00001442
1443<p>Consequently, type-based alias analysis, aka TBAA, aka
1444<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1445LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1446additional information which specialized optimization passes may use
1447to implement type-based alias analysis.</p>
1448
1449</div>
1450
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001451<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001452<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001453 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001454</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001455
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001456<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001457
1458<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1459href="#i_store"><tt>store</tt></a>s, and <a
1460href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1461The optimizers must not change the number of volatile operations or change their
1462order of execution relative to other volatile operations. The optimizers
1463<i>may</i> change the order of volatile operations relative to non-volatile
1464operations. This is not Java's "volatile" and has no cross-thread
1465synchronization behavior.</p>
1466
1467</div>
1468
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001469</div>
1470
Chris Lattner00950542001-06-06 20:29:01 +00001471<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001472<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +00001473<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +00001474
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001475<div>
Chris Lattnerfa730212004-12-09 16:11:40 +00001476
Misha Brukman9d0919f2003-11-08 01:05:38 +00001477<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001478 intermediate representation. Being typed enables a number of optimizations
1479 to be performed on the intermediate representation directly, without having
1480 to do extra analyses on the side before the transformation. A strong type
1481 system makes it easier to read the generated code and enables novel analyses
1482 and transformations that are not feasible to perform on normal three address
1483 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +00001484
Chris Lattner00950542001-06-06 20:29:01 +00001485<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001486<h3>
1487 <a name="t_classifications">Type Classifications</a>
1488</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001489
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001490<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001491
1492<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001493
1494<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001495 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001496 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +00001497 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001498 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +00001499 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +00001500 </tr>
1501 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001502 <td><a href="#t_floating">floating point</a></td>
1503 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +00001504 </tr>
1505 <tr>
1506 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +00001507 <td><a href="#t_integer">integer</a>,
1508 <a href="#t_floating">floating point</a>,
1509 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +00001510 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +00001511 <a href="#t_struct">structure</a>,
1512 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001513 <a href="#t_label">label</a>,
1514 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +00001515 </td>
Chris Lattner261efe92003-11-25 01:02:51 +00001516 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001517 <tr>
1518 <td><a href="#t_primitive">primitive</a></td>
1519 <td><a href="#t_label">label</a>,
1520 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +00001521 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001522 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001523 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001524 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +00001525 </tr>
1526 <tr>
1527 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001528 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +00001529 <a href="#t_function">function</a>,
1530 <a href="#t_pointer">pointer</a>,
1531 <a href="#t_struct">structure</a>,
1532 <a href="#t_pstruct">packed structure</a>,
1533 <a href="#t_vector">vector</a>,
1534 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +00001535 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +00001536 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +00001537 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001538</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001539
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001540<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1541 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +00001542 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001543
Misha Brukman9d0919f2003-11-08 01:05:38 +00001544</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001545
Chris Lattner00950542001-06-06 20:29:01 +00001546<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001547<h3>
1548 <a name="t_primitive">Primitive Types</a>
1549</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +00001550
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001551<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001552
Chris Lattner4f69f462008-01-04 04:32:38 +00001553<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001554 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +00001555
1556<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001557<h4>
1558 <a name="t_integer">Integer Type</a>
1559</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +00001560
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001561<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +00001562
1563<h5>Overview:</h5>
1564<p>The integer type is a very simple type that simply specifies an arbitrary
1565 bit width for the integer type desired. Any bit width from 1 bit to
1566 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1567
1568<h5>Syntax:</h5>
1569<pre>
1570 iN
1571</pre>
1572
1573<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1574 value.</p>
1575
1576<h5>Examples:</h5>
1577<table class="layout">
1578 <tr class="layout">
1579 <td class="left"><tt>i1</tt></td>
1580 <td class="left">a single-bit integer.</td>
1581 </tr>
1582 <tr class="layout">
1583 <td class="left"><tt>i32</tt></td>
1584 <td class="left">a 32-bit integer.</td>
1585 </tr>
1586 <tr class="layout">
1587 <td class="left"><tt>i1942652</tt></td>
1588 <td class="left">a really big integer of over 1 million bits.</td>
1589 </tr>
1590</table>
1591
Nick Lewyckyec38da42009-09-27 00:45:11 +00001592</div>
1593
1594<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001595<h4>
1596 <a name="t_floating">Floating Point Types</a>
1597</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +00001598
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001599<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001600
1601<table>
1602 <tbody>
1603 <tr><th>Type</th><th>Description</th></tr>
1604 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1605 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1606 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1607 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1608 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1609 </tbody>
1610</table>
1611
Chris Lattner4f69f462008-01-04 04:32:38 +00001612</div>
1613
1614<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001615<h4>
1616 <a name="t_x86mmx">X86mmx Type</a>
1617</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001618
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001619<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001620
1621<h5>Overview:</h5>
1622<p>The x86mmx type represents a value held in an MMX register on an x86 machine. The operations allowed on it are quite limited: parameters and return values, load and store, and bitcast. User-specified MMX instructions are represented as intrinsic or asm calls with arguments and/or results of this type. There are no arrays, vectors or constants of this type.</p>
1623
1624<h5>Syntax:</h5>
1625<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +00001626 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001627</pre>
1628
1629</div>
1630
1631<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001632<h4>
1633 <a name="t_void">Void Type</a>
1634</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +00001635
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001636<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001637
Chris Lattner4f69f462008-01-04 04:32:38 +00001638<h5>Overview:</h5>
1639<p>The void type does not represent any value and has no size.</p>
1640
1641<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +00001642<pre>
1643 void
1644</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001645
Chris Lattner4f69f462008-01-04 04:32:38 +00001646</div>
1647
1648<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001649<h4>
1650 <a name="t_label">Label Type</a>
1651</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +00001652
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001653<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001654
Chris Lattner4f69f462008-01-04 04:32:38 +00001655<h5>Overview:</h5>
1656<p>The label type represents code labels.</p>
1657
1658<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +00001659<pre>
1660 label
1661</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001662
Chris Lattner4f69f462008-01-04 04:32:38 +00001663</div>
1664
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001665<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001666<h4>
1667 <a name="t_metadata">Metadata Type</a>
1668</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001669
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001670<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001671
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001672<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +00001673<p>The metadata type represents embedded metadata. No derived types may be
1674 created from metadata except for <a href="#t_function">function</a>
1675 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001676
1677<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001678<pre>
1679 metadata
1680</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001681
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001682</div>
1683
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001684</div>
Chris Lattner4f69f462008-01-04 04:32:38 +00001685
1686<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001687<h3>
1688 <a name="t_derived">Derived Types</a>
1689</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001690
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001691<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001692
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001693<p>The real power in LLVM comes from the derived types in the system. This is
1694 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +00001695 useful types. Each of these types contain one or more element types which
1696 may be a primitive type, or another derived type. For example, it is
1697 possible to have a two dimensional array, using an array as the element type
1698 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +00001699
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001700
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001701<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001702<h4>
1703 <a name="t_aggregate">Aggregate Types</a>
1704</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001705
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001706<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001707
1708<p>Aggregate Types are a subset of derived types that can contain multiple
1709 member types. <a href="#t_array">Arrays</a>,
Chris Lattner61c70e92010-08-28 04:09:24 +00001710 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1711 aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001712
1713</div>
1714
Reid Spencer2b916312007-05-16 18:44:01 +00001715<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001716<h4>
1717 <a name="t_array">Array Type</a>
1718</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001719
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001720<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001721
Chris Lattner00950542001-06-06 20:29:01 +00001722<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001723<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001724 sequentially in memory. The array type requires a size (number of elements)
1725 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001726
Chris Lattner7faa8832002-04-14 06:13:44 +00001727<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001728<pre>
1729 [&lt;# elements&gt; x &lt;elementtype&gt;]
1730</pre>
1731
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001732<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1733 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001734
Chris Lattner7faa8832002-04-14 06:13:44 +00001735<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001736<table class="layout">
1737 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001738 <td class="left"><tt>[40 x i32]</tt></td>
1739 <td class="left">Array of 40 32-bit integer values.</td>
1740 </tr>
1741 <tr class="layout">
1742 <td class="left"><tt>[41 x i32]</tt></td>
1743 <td class="left">Array of 41 32-bit integer values.</td>
1744 </tr>
1745 <tr class="layout">
1746 <td class="left"><tt>[4 x i8]</tt></td>
1747 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001748 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001749</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001750<p>Here are some examples of multidimensional arrays:</p>
1751<table class="layout">
1752 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001753 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1754 <td class="left">3x4 array of 32-bit integer values.</td>
1755 </tr>
1756 <tr class="layout">
1757 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1758 <td class="left">12x10 array of single precision floating point values.</td>
1759 </tr>
1760 <tr class="layout">
1761 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1762 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001763 </tr>
1764</table>
Chris Lattnere67a9512005-06-24 17:22:57 +00001765
Dan Gohman7657f6b2009-11-09 19:01:53 +00001766<p>There is no restriction on indexing beyond the end of the array implied by
1767 a static type (though there are restrictions on indexing beyond the bounds
1768 of an allocated object in some cases). This means that single-dimension
1769 'variable sized array' addressing can be implemented in LLVM with a zero
1770 length array type. An implementation of 'pascal style arrays' in LLVM could
1771 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +00001772
Misha Brukman9d0919f2003-11-08 01:05:38 +00001773</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001774
Chris Lattner00950542001-06-06 20:29:01 +00001775<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001776<h4>
1777 <a name="t_function">Function Type</a>
1778</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001779
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001780<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001781
Chris Lattner00950542001-06-06 20:29:01 +00001782<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001783<p>The function type can be thought of as a function signature. It consists of
1784 a return type and a list of formal parameter types. The return type of a
Chris Lattner61c70e92010-08-28 04:09:24 +00001785 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001786
Chris Lattner00950542001-06-06 20:29:01 +00001787<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001788<pre>
Nick Lewycky51386942009-09-27 07:55:32 +00001789 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001790</pre>
1791
John Criswell0ec250c2005-10-24 16:17:18 +00001792<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001793 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1794 which indicates that the function takes a variable number of arguments.
1795 Variable argument functions can access their arguments with
1796 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +00001797 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +00001798 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001799
Chris Lattner00950542001-06-06 20:29:01 +00001800<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001801<table class="layout">
1802 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +00001803 <td class="left"><tt>i32 (i32)</tt></td>
1804 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001805 </td>
Reid Spencer92f82302006-12-31 07:18:34 +00001806 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +00001807 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +00001808 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001809 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +00001810 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1811 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +00001812 </td>
1813 </tr><tr class="layout">
1814 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001815 <td class="left">A vararg function that takes at least one
1816 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1817 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +00001818 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +00001819 </td>
Devang Patela582f402008-03-24 05:35:41 +00001820 </tr><tr class="layout">
1821 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +00001822 <td class="left">A function taking an <tt>i32</tt>, returning a
1823 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +00001824 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001825 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001826</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001827
Misha Brukman9d0919f2003-11-08 01:05:38 +00001828</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001829
Chris Lattner00950542001-06-06 20:29:01 +00001830<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001831<h4>
1832 <a name="t_struct">Structure Type</a>
1833</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001834
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001835<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001836
Chris Lattner00950542001-06-06 20:29:01 +00001837<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001838<p>The structure type is used to represent a collection of data members together
1839 in memory. The packing of the field types is defined to match the ABI of the
1840 underlying processor. The elements of a structure may be any type that has a
1841 size.</p>
1842
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +00001843<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1844 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1845 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1846 Structures in registers are accessed using the
1847 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1848 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner00950542001-06-06 20:29:01 +00001849<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001850<pre>
1851 { &lt;type list&gt; }
1852</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001853
Chris Lattner00950542001-06-06 20:29:01 +00001854<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001855<table class="layout">
1856 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001857 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1858 <td class="left">A triple of three <tt>i32</tt> values</td>
1859 </tr><tr class="layout">
1860 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1861 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1862 second element is a <a href="#t_pointer">pointer</a> to a
1863 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1864 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001865 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001866</table>
Dan Gohmand8791e52009-01-24 15:58:40 +00001867
Misha Brukman9d0919f2003-11-08 01:05:38 +00001868</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001869
Chris Lattner00950542001-06-06 20:29:01 +00001870<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001871<h4>
1872 <a name="t_pstruct">Packed Structure Type</a>
1873</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001874
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001875<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001876
Andrew Lenharth75e10682006-12-08 17:13:00 +00001877<h5>Overview:</h5>
1878<p>The packed structure type is used to represent a collection of data members
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001879 together in memory. There is no padding between fields. Further, the
1880 alignment of a packed structure is 1 byte. The elements of a packed
1881 structure may be any type that has a size.</p>
1882
1883<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1884 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1885 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1886
Andrew Lenharth75e10682006-12-08 17:13:00 +00001887<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001888<pre>
1889 &lt; { &lt;type list&gt; } &gt;
1890</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001891
Andrew Lenharth75e10682006-12-08 17:13:00 +00001892<h5>Examples:</h5>
1893<table class="layout">
1894 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001895 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1896 <td class="left">A triple of three <tt>i32</tt> values</td>
1897 </tr><tr class="layout">
Bill Wendlinge36dccc2008-09-07 10:26:33 +00001898 <td class="left">
1899<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001900 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1901 second element is a <a href="#t_pointer">pointer</a> to a
1902 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1903 an <tt>i32</tt>.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +00001904 </tr>
1905</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001906
Andrew Lenharth75e10682006-12-08 17:13:00 +00001907</div>
1908
1909<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001910<h4>
1911 <a name="t_pointer">Pointer Type</a>
1912</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +00001913
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001914<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001915
1916<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +00001917<p>The pointer type is used to specify memory locations.
1918 Pointers are commonly used to reference objects in memory.</p>
1919
1920<p>Pointer types may have an optional address space attribute defining the
1921 numbered address space where the pointed-to object resides. The default
1922 address space is number zero. The semantics of non-zero address
1923 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001924
1925<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1926 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +00001927
Chris Lattner7faa8832002-04-14 06:13:44 +00001928<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001929<pre>
1930 &lt;type&gt; *
1931</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001932
Chris Lattner7faa8832002-04-14 06:13:44 +00001933<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001934<table class="layout">
1935 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +00001936 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +00001937 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1938 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1939 </tr>
1940 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +00001941 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +00001942 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +00001943 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +00001944 <tt>i32</tt>.</td>
1945 </tr>
1946 <tr class="layout">
1947 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1948 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1949 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001950 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001951</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001952
Misha Brukman9d0919f2003-11-08 01:05:38 +00001953</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001954
Chris Lattnera58561b2004-08-12 19:12:28 +00001955<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001956<h4>
1957 <a name="t_vector">Vector Type</a>
1958</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001959
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001960<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001961
Chris Lattnera58561b2004-08-12 19:12:28 +00001962<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001963<p>A vector type is a simple derived type that represents a vector of elements.
1964 Vector types are used when multiple primitive data are operated in parallel
1965 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +00001966 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001967 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001968
Chris Lattnera58561b2004-08-12 19:12:28 +00001969<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001970<pre>
1971 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1972</pre>
1973
Chris Lattner7d2e7be2010-10-10 18:20:35 +00001974<p>The number of elements is a constant integer value larger than 0; elementtype
1975 may be any integer or floating point type. Vectors of size zero are not
1976 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +00001977
Chris Lattnera58561b2004-08-12 19:12:28 +00001978<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001979<table class="layout">
1980 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001981 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1982 <td class="left">Vector of 4 32-bit integer values.</td>
1983 </tr>
1984 <tr class="layout">
1985 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1986 <td class="left">Vector of 8 32-bit floating-point values.</td>
1987 </tr>
1988 <tr class="layout">
1989 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1990 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001991 </tr>
1992</table>
Dan Gohmand8791e52009-01-24 15:58:40 +00001993
Misha Brukman9d0919f2003-11-08 01:05:38 +00001994</div>
1995
Chris Lattner69c11bb2005-04-25 17:34:15 +00001996<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001997<h4>
1998 <a name="t_opaque">Opaque Type</a>
1999</h4>
2000
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002001<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002002
2003<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002004<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002005 corresponds (for example) to the C notion of a forward declared structure
2006 type. In LLVM, opaque types can eventually be resolved to any type (not just
2007 a structure type).</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002008
2009<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002010<pre>
2011 opaque
2012</pre>
2013
2014<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002015<table class="layout">
2016 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00002017 <td class="left"><tt>opaque</tt></td>
2018 <td class="left">An opaque type.</td>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002019 </tr>
2020</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002021
Chris Lattner69c11bb2005-04-25 17:34:15 +00002022</div>
2023
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002024</div>
2025
Chris Lattner242d61d2009-02-02 07:32:36 +00002026<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002027<h3>
Chris Lattner242d61d2009-02-02 07:32:36 +00002028 <a name="t_uprefs">Type Up-references</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002029</h3>
Chris Lattner242d61d2009-02-02 07:32:36 +00002030
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002031<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002032
Chris Lattner242d61d2009-02-02 07:32:36 +00002033<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002034<p>An "up reference" allows you to refer to a lexically enclosing type without
2035 requiring it to have a name. For instance, a structure declaration may
2036 contain a pointer to any of the types it is lexically a member of. Example
2037 of up references (with their equivalent as named type declarations)
2038 include:</p>
Chris Lattner242d61d2009-02-02 07:32:36 +00002039
2040<pre>
Chris Lattner3060f5b2009-02-09 10:00:56 +00002041 { \2 * } %x = type { %x* }
Chris Lattner242d61d2009-02-02 07:32:36 +00002042 { \2 }* %y = type { %y }*
2043 \1* %z = type %z*
2044</pre>
2045
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002046<p>An up reference is needed by the asmprinter for printing out cyclic types
2047 when there is no declared name for a type in the cycle. Because the
2048 asmprinter does not want to print out an infinite type string, it needs a
2049 syntax to handle recursive types that have no names (all names are optional
2050 in llvm IR).</p>
Chris Lattner242d61d2009-02-02 07:32:36 +00002051
2052<h5>Syntax:</h5>
2053<pre>
2054 \&lt;level&gt;
2055</pre>
2056
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002057<p>The level is the count of the lexical type that is being referred to.</p>
Chris Lattner242d61d2009-02-02 07:32:36 +00002058
2059<h5>Examples:</h5>
Chris Lattner242d61d2009-02-02 07:32:36 +00002060<table class="layout">
2061 <tr class="layout">
2062 <td class="left"><tt>\1*</tt></td>
2063 <td class="left">Self-referential pointer.</td>
2064 </tr>
2065 <tr class="layout">
2066 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
2067 <td class="left">Recursive structure where the upref refers to the out-most
2068 structure.</td>
2069 </tr>
2070</table>
Chris Lattner242d61d2009-02-02 07:32:36 +00002071
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002072</div>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002073
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002074</div>
2075
Chris Lattnerc3f59762004-12-09 17:30:23 +00002076<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002077<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002078<!-- *********************************************************************** -->
2079
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002080<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002081
2082<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002083 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002084
Chris Lattnerc3f59762004-12-09 17:30:23 +00002085<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002086<h3>
2087 <a name="simpleconstants">Simple Constants</a>
2088</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002089
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002090<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002091
2092<dl>
2093 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002094 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +00002095 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002096
2097 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002098 <dd>Standard integers (such as '4') are constants of
2099 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2100 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002101
2102 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002103 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002104 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2105 notation (see below). The assembler requires the exact decimal value of a
2106 floating-point constant. For example, the assembler accepts 1.25 but
2107 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2108 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002109
2110 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +00002111 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002112 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002113</dl>
2114
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002115<p>The one non-intuitive notation for constants is the hexadecimal form of
2116 floating point constants. For example, the form '<tt>double
2117 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2118 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2119 constants are required (and the only time that they are generated by the
2120 disassembler) is when a floating point constant must be emitted but it cannot
2121 be represented as a decimal floating point number in a reasonable number of
2122 digits. For example, NaN's, infinities, and other special values are
2123 represented in their IEEE hexadecimal format so that assembly and disassembly
2124 do not cause any bits to change in the constants.</p>
2125
Dale Johannesenbd5e5a82009-02-11 22:14:51 +00002126<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002127 represented using the 16-digit form shown above (which matches the IEEE754
2128 representation for double); float values must, however, be exactly
2129 representable as IEE754 single precision. Hexadecimal format is always used
2130 for long double, and there are three forms of long double. The 80-bit format
2131 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2132 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2133 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2134 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2135 currently supported target uses this format. Long doubles will only work if
2136 they match the long double format on your target. All hexadecimal formats
2137 are big-endian (sign bit at the left).</p>
2138
Dale Johannesen21fe99b2010-10-01 00:48:59 +00002139<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002140</div>
2141
2142<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002143<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +00002144<a name="aggregateconstants"></a> <!-- old anchor -->
2145<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002146</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002147
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002148<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002149
Chris Lattner70882792009-02-28 18:32:25 +00002150<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002151 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002152
2153<dl>
2154 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002155 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002156 type definitions (a comma separated list of elements, surrounded by braces
2157 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2158 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2159 Structure constants must have <a href="#t_struct">structure type</a>, and
2160 the number and types of elements must match those specified by the
2161 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002162
2163 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002164 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002165 definitions (a comma separated list of elements, surrounded by square
2166 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2167 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2168 the number and types of elements must match those specified by the
2169 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002170
Reid Spencer485bad12007-02-15 03:07:05 +00002171 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +00002172 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002173 definitions (a comma separated list of elements, surrounded by
2174 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2175 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2176 have <a href="#t_vector">vector type</a>, and the number and types of
2177 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002178
2179 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002180 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +00002181 value to zero of <em>any</em> type, including scalar and
2182 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002183 This is often used to avoid having to print large zero initializers
2184 (e.g. for large arrays) and is always exactly equivalent to using explicit
2185 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +00002186
2187 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +00002188 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002189 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2190 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2191 be interpreted as part of the instruction stream, metadata is a place to
2192 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002193</dl>
2194
2195</div>
2196
2197<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002198<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002199 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002200</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002201
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002202<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002203
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002204<p>The addresses of <a href="#globalvars">global variables</a>
2205 and <a href="#functionstructure">functions</a> are always implicitly valid
2206 (link-time) constants. These constants are explicitly referenced when
2207 the <a href="#identifiers">identifier for the global</a> is used and always
2208 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2209 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002210
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002211<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +00002212@X = global i32 17
2213@Y = global i32 42
2214@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +00002215</pre>
2216
2217</div>
2218
2219<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002220<h3>
2221 <a name="undefvalues">Undefined Values</a>
2222</h3>
2223
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002224<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002225
Chris Lattner48a109c2009-09-07 22:52:39 +00002226<p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
Benjamin Kramer8040cd32009-10-12 14:46:08 +00002227 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +00002228 Undefined values may be of any type (other than '<tt>label</tt>'
2229 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002230
Chris Lattnerc608cb12009-09-11 01:49:31 +00002231<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +00002232 program is well defined no matter what value is used. This gives the
2233 compiler more freedom to optimize. Here are some examples of (potentially
2234 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002235
Chris Lattner48a109c2009-09-07 22:52:39 +00002236
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002237<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002238 %A = add %X, undef
2239 %B = sub %X, undef
2240 %C = xor %X, undef
2241Safe:
2242 %A = undef
2243 %B = undef
2244 %C = undef
2245</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002246
2247<p>This is safe because all of the output bits are affected by the undef bits.
Bill Wendling1b383ba2010-10-27 01:07:41 +00002248 Any output bit can have a zero or one depending on the input bits.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +00002249
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002250<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002251 %A = or %X, undef
2252 %B = and %X, undef
2253Safe:
2254 %A = -1
2255 %B = 0
2256Unsafe:
2257 %A = undef
2258 %B = undef
2259</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002260
2261<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +00002262 For example, if <tt>%X</tt> has a zero bit, then the output of the
2263 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2264 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2265 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2266 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2267 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2268 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2269 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +00002270
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002271<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002272 %A = select undef, %X, %Y
2273 %B = select undef, 42, %Y
2274 %C = select %X, %Y, undef
2275Safe:
2276 %A = %X (or %Y)
2277 %B = 42 (or %Y)
2278 %C = %Y
2279Unsafe:
2280 %A = undef
2281 %B = undef
2282 %C = undef
2283</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002284
Bill Wendling1b383ba2010-10-27 01:07:41 +00002285<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2286 branch) conditions can go <em>either way</em>, but they have to come from one
2287 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2288 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2289 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2290 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2291 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2292 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +00002293
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002294<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002295 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002296
Chris Lattner48a109c2009-09-07 22:52:39 +00002297 %B = undef
2298 %C = xor %B, %B
2299
2300 %D = undef
2301 %E = icmp lt %D, 4
2302 %F = icmp gte %D, 4
2303
2304Safe:
2305 %A = undef
2306 %B = undef
2307 %C = undef
2308 %D = undef
2309 %E = undef
2310 %F = undef
2311</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002312
Bill Wendling1b383ba2010-10-27 01:07:41 +00002313<p>This example points out that two '<tt>undef</tt>' operands are not
2314 necessarily the same. This can be surprising to people (and also matches C
2315 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2316 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2317 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2318 its value over its "live range". This is true because the variable doesn't
2319 actually <em>have a live range</em>. Instead, the value is logically read
2320 from arbitrary registers that happen to be around when needed, so the value
2321 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2322 need to have the same semantics or the core LLVM "replace all uses with"
2323 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002324
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002325<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +00002326 %A = fdiv undef, %X
2327 %B = fdiv %X, undef
2328Safe:
2329 %A = undef
2330b: unreachable
2331</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002332
2333<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +00002334 value</em> and <em>undefined behavior</em>. An undefined value (like
2335 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2336 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2337 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2338 defined on SNaN's. However, in the second example, we can make a more
2339 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2340 arbitrary value, we are allowed to assume that it could be zero. Since a
2341 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2342 the operation does not execute at all. This allows us to delete the divide and
2343 all code after it. Because the undefined operation "can't happen", the
2344 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002345
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002346<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +00002347a: store undef -> %X
2348b: store %X -> undef
2349Safe:
2350a: &lt;deleted&gt;
2351b: unreachable
2352</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002353
Bill Wendling1b383ba2010-10-27 01:07:41 +00002354<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2355 undefined value can be assumed to not have any effect; we can assume that the
2356 value is overwritten with bits that happen to match what was already there.
2357 However, a store <em>to</em> an undefined location could clobber arbitrary
2358 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002359
Chris Lattnerc3f59762004-12-09 17:30:23 +00002360</div>
2361
2362<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002363<h3>
2364 <a name="trapvalues">Trap Values</a>
2365</h3>
2366
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002367<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002368
Dan Gohmanc68ce062010-04-26 20:21:21 +00002369<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanfff6c532010-04-22 23:14:21 +00002370 instead of representing an unspecified bit pattern, they represent the
2371 fact that an instruction or constant expression which cannot evoke side
2372 effects has nevertheless detected a condition which results in undefined
Dan Gohmanc68ce062010-04-26 20:21:21 +00002373 behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002374
Dan Gohman34b3d992010-04-28 00:49:41 +00002375<p>There is currently no way of representing a trap value in the IR; they
Dan Gohman855abed2010-05-03 14:51:43 +00002376 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +00002377 <a href="#i_add"><tt>add</tt></a> with the <tt>nsw</tt> flag.</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002378
Dan Gohman34b3d992010-04-28 00:49:41 +00002379<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002380
Dan Gohman34b3d992010-04-28 00:49:41 +00002381<ul>
2382<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2383 their operands.</li>
2384
2385<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2386 to their dynamic predecessor basic block.</li>
2387
2388<li>Function arguments depend on the corresponding actual argument values in
2389 the dynamic callers of their functions.</li>
2390
2391<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2392 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2393 control back to them.</li>
2394
Dan Gohmanb5328162010-05-03 14:55:22 +00002395<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2396 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2397 or exception-throwing call instructions that dynamically transfer control
2398 back to them.</li>
2399
Dan Gohman34b3d992010-04-28 00:49:41 +00002400<li>Non-volatile loads and stores depend on the most recent stores to all of the
2401 referenced memory addresses, following the order in the IR
2402 (including loads and stores implied by intrinsics such as
2403 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2404
Dan Gohman7c24ff12010-05-03 14:59:34 +00002405<!-- TODO: In the case of multiple threads, this only applies if the store
2406 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002407
Dan Gohman34b3d992010-04-28 00:49:41 +00002408<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002409
Dan Gohman34b3d992010-04-28 00:49:41 +00002410<li>An instruction with externally visible side effects depends on the most
2411 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +00002412 the order in the IR. (This includes
2413 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002414
Dan Gohmanb5328162010-05-03 14:55:22 +00002415<li>An instruction <i>control-depends</i> on a
2416 <a href="#terminators">terminator instruction</a>
2417 if the terminator instruction has multiple successors and the instruction
2418 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +00002419 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +00002420
Dan Gohmanca4cac42011-04-12 23:05:59 +00002421<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2422 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +00002423 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +00002424 successor.</li>
2425
Dan Gohman34b3d992010-04-28 00:49:41 +00002426<li>Dependence is transitive.</li>
2427
2428</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +00002429
2430<p>Whenever a trap value is generated, all values which depend on it evaluate
2431 to trap. If they have side effects, the evoke their side effects as if each
2432 operand with a trap value were undef. If they have externally-visible side
2433 effects, the behavior is undefined.</p>
2434
2435<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +00002436
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002437<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002438entry:
2439 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +00002440 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2441 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2442 store i32 0, i32* %trap_yet_again ; undefined behavior
2443
2444 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2445 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2446
2447 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2448
2449 %narrowaddr = bitcast i32* @g to i16*
2450 %wideaddr = bitcast i32* @g to i64*
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002451 %trap3 = load i16* %narrowaddr ; Returns a trap value.
2452 %trap4 = load i64* %wideaddr ; Returns a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +00002453
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002454 %cmp = icmp slt i32 %trap, 0 ; Returns a trap value.
2455 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002456
2457true:
Dan Gohman34b3d992010-04-28 00:49:41 +00002458 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2459 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002460 br label %end
2461
2462end:
2463 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2464 ; Both edges into this PHI are
2465 ; control-dependent on %cmp, so this
Dan Gohman34b3d992010-04-28 00:49:41 +00002466 ; always results in a trap value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002467
Dan Gohmanca4cac42011-04-12 23:05:59 +00002468 volatile store i32 0, i32* @g ; This would depend on the store in %true
2469 ; if %cmp is true, or the store in %entry
2470 ; otherwise, so this is undefined behavior.
2471
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002472 br i1 %cmp, label %second_true, label %second_end
Dan Gohmanca4cac42011-04-12 23:05:59 +00002473 ; The same branch again, but this time the
2474 ; true block doesn't have side effects.
2475
2476second_true:
2477 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002478 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +00002479
2480second_end:
2481 volatile store i32 0, i32* @g ; This time, the instruction always depends
2482 ; on the store in %end. Also, it is
2483 ; control-equivalent to %end, so this is
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002484 ; well-defined (again, ignoring earlier
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002485 ; undefined behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002486</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002487
Dan Gohmanfff6c532010-04-22 23:14:21 +00002488</div>
2489
2490<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002491<h3>
2492 <a name="blockaddress">Addresses of Basic Blocks</a>
2493</h3>
2494
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002495<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +00002496
Chris Lattnercdfc9402009-11-01 01:27:45 +00002497<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +00002498
2499<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +00002500 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +00002501 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002502
Chris Lattnerc6f44362009-10-27 21:01:34 +00002503<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +00002504 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2505 comparisons against null. Pointer equality tests between labels addresses
2506 results in undefined behavior &mdash; though, again, comparison against null
2507 is ok, and no label is equal to the null pointer. This may be passed around
2508 as an opaque pointer sized value as long as the bits are not inspected. This
2509 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2510 long as the original value is reconstituted before the <tt>indirectbr</tt>
2511 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002512
Bill Wendling1b383ba2010-10-27 01:07:41 +00002513<p>Finally, some targets may provide defined semantics when using the value as
2514 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +00002515
2516</div>
2517
2518
2519<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002520<h3>
2521 <a name="constantexprs">Constant Expressions</a>
2522</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002523
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002524<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002525
2526<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002527 to be used as constants. Constant expressions may be of
2528 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2529 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +00002530 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002531
2532<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002533 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002534 <dd>Truncate a constant to another type. The bit size of CST must be larger
2535 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002536
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002537 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002538 <dd>Zero extend a constant to another type. The bit size of CST must be
Duncan Sands28afd432010-07-13 12:06:14 +00002539 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002540
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002541 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002542 <dd>Sign extend a constant to another type. The bit size of CST must be
Duncan Sands28afd432010-07-13 12:06:14 +00002543 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002544
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002545 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002546 <dd>Truncate a floating point constant to another floating point type. The
2547 size of CST must be larger than the size of TYPE. Both types must be
2548 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002549
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002550 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002551 <dd>Floating point extend a constant to another type. The size of CST must be
2552 smaller or equal to the size of TYPE. Both types must be floating
2553 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002554
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002555 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002556 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002557 constant. TYPE must be a scalar or vector integer type. CST must be of
2558 scalar or vector floating point type. Both CST and TYPE must be scalars,
2559 or vectors of the same number of elements. If the value won't fit in the
2560 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002561
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002562 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002563 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002564 constant. TYPE must be a scalar or vector integer type. CST must be of
2565 scalar or vector floating point type. Both CST and TYPE must be scalars,
2566 or vectors of the same number of elements. If the value won't fit in the
2567 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002568
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002569 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002570 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002571 constant. TYPE must be a scalar or vector floating point type. CST must be
2572 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2573 vectors of the same number of elements. If the value won't fit in the
2574 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002575
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002576 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002577 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002578 constant. TYPE must be a scalar or vector floating point type. CST must be
2579 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2580 vectors of the same number of elements. If the value won't fit in the
2581 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002582
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002583 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +00002584 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002585 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2586 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2587 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +00002588
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002589 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002590 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2591 type. CST must be of integer type. The CST value is zero extended,
2592 truncated, or unchanged to make it fit in a pointer size. This one is
2593 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +00002594
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002595 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +00002596 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2597 are the same as those for the <a href="#i_bitcast">bitcast
2598 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002599
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002600 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2601 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002602 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002603 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2604 instruction, the index list may have zero or more indexes, which are
2605 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002606
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002607 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002608 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +00002609
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002610 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +00002611 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2612
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002613 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +00002614 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +00002615
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002616 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002617 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2618 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +00002619
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002620 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002621 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2622 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +00002623
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002624 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002625 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2626 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +00002627
Nick Lewycky9e130ce2010-05-29 06:44:15 +00002628 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2629 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2630 constants. The index list is interpreted in a similar manner as indices in
2631 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2632 index value must be specified.</dd>
2633
2634 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2635 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2636 constants. The index list is interpreted in a similar manner as indices in
2637 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2638 index value must be specified.</dd>
2639
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002640 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002641 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2642 be any of the <a href="#binaryops">binary</a>
2643 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2644 on operands are the same as those for the corresponding instruction
2645 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002646</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002647
Chris Lattnerc3f59762004-12-09 17:30:23 +00002648</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00002649
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002650</div>
2651
Chris Lattner00950542001-06-06 20:29:01 +00002652<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002653<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +00002654<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002655<div>
Chris Lattnere87d6532006-01-25 23:47:57 +00002656<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002657<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +00002658<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002659</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +00002660
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002661<div>
Chris Lattnere87d6532006-01-25 23:47:57 +00002662
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002663<p>LLVM supports inline assembler expressions (as opposed
2664 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2665 a special value. This value represents the inline assembler as a string
2666 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +00002667 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +00002668 expression has side effects, and a flag indicating whether the function
2669 containing the asm needs to align its stack conservatively. An example
2670 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002671
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002672<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00002673i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +00002674</pre>
2675
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002676<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2677 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2678 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002679
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002680<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00002681%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattnere87d6532006-01-25 23:47:57 +00002682</pre>
2683
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002684<p>Inline asms with side effects not visible in the constraint list must be
2685 marked as having side effects. This is done through the use of the
2686 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002687
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002688<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00002689call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +00002690</pre>
2691
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +00002692<p>In some cases inline asms will contain code that will not work unless the
2693 stack is aligned in some way, such as calls or SSE instructions on x86,
2694 yet will not contain code that does that alignment within the asm.
2695 The compiler should make conservative assumptions about what the asm might
2696 contain and should generate its usual stack alignment code in the prologue
2697 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +00002698
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002699<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +00002700call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +00002701</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +00002702
2703<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2704 first.</p>
2705
Chris Lattnere87d6532006-01-25 23:47:57 +00002706<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002707 documented here. Constraints on what can be done (e.g. duplication, moving,
2708 etc need to be documented). This is probably best done by reference to
2709 another document that covers inline asm from a holistic perspective.</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002710
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002711<h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002712<a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002713</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002714
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002715<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002716
2717<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattnerce1b9ad2010-11-17 08:20:42 +00002718 attached to it that contains a list of constant integers. If present, the
2719 code generator will use the integer as the location cookie value when report
Chris Lattnercf9a4152010-04-07 05:38:05 +00002720 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman1c70c002010-04-28 00:36:01 +00002721 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattnercf9a4152010-04-07 05:38:05 +00002722 source code that produced it. For example:</p>
2723
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002724<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +00002725call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2726...
2727!42 = !{ i32 1234567 }
2728</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002729
2730<p>It is up to the front-end to make sense of the magic numbers it places in the
Chris Lattnerce1b9ad2010-11-17 08:20:42 +00002731 IR. If the MDNode contains multiple constants, the code generator will use
2732 the one that corresponds to the line of the asm that the error occurs on.</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002733
2734</div>
2735
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002736</div>
2737
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002738<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002739<h3>
2740 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2741</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002742
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002743<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002744
2745<p>LLVM IR allows metadata to be attached to instructions in the program that
2746 can convey extra information about the code to the optimizers and code
2747 generator. One example application of metadata is source-level debug
2748 information. There are two metadata primitives: strings and nodes. All
2749 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2750 preceding exclamation point ('<tt>!</tt>').</p>
2751
2752<p>A metadata string is a string surrounded by double quotes. It can contain
2753 any character by escaping non-printable characters with "\xx" where "xx" is
2754 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2755
2756<p>Metadata nodes are represented with notation similar to structure constants
2757 (a comma separated list of elements, surrounded by braces and preceded by an
2758 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2759 10}</tt>". Metadata nodes can have any values as their operand.</p>
2760
2761<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2762 metadata nodes, which can be looked up in the module symbol table. For
2763 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2764
Devang Patele1d50cd2010-03-04 23:44:48 +00002765<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002766 function is using two metadata arguments.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +00002767
Bill Wendling9ff5de92011-03-02 02:17:11 +00002768<div class="doc_code">
2769<pre>
2770call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2771</pre>
2772</div>
Devang Patele1d50cd2010-03-04 23:44:48 +00002773
2774<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002775 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +00002776
Bill Wendling9ff5de92011-03-02 02:17:11 +00002777<div class="doc_code">
2778<pre>
2779%indvar.next = add i64 %indvar, 1, !dbg !21
2780</pre>
2781</div>
2782
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002783</div>
2784
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002785</div>
Chris Lattner857755c2009-07-20 05:55:19 +00002786
2787<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002788<h2>
Chris Lattner857755c2009-07-20 05:55:19 +00002789 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002790</h2>
Chris Lattner857755c2009-07-20 05:55:19 +00002791<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002792<div>
Chris Lattner857755c2009-07-20 05:55:19 +00002793<p>LLVM has a number of "magic" global variables that contain data that affect
2794code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +00002795of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2796section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2797by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +00002798
2799<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002800<h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002801<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002802</h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002803
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002804<div>
Chris Lattner857755c2009-07-20 05:55:19 +00002805
2806<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2807href="#linkage_appending">appending linkage</a>. This array contains a list of
2808pointers to global variables and functions which may optionally have a pointer
2809cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2810
2811<pre>
2812 @X = global i8 4
2813 @Y = global i32 123
2814
2815 @llvm.used = appending global [2 x i8*] [
2816 i8* @X,
2817 i8* bitcast (i32* @Y to i8*)
2818 ], section "llvm.metadata"
2819</pre>
2820
2821<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2822compiler, assembler, and linker are required to treat the symbol as if there is
2823a reference to the global that it cannot see. For example, if a variable has
2824internal linkage and no references other than that from the <tt>@llvm.used</tt>
2825list, it cannot be deleted. This is commonly used to represent references from
2826inline asms and other things the compiler cannot "see", and corresponds to
2827"attribute((used))" in GNU C.</p>
2828
2829<p>On some targets, the code generator must emit a directive to the assembler or
2830object file to prevent the assembler and linker from molesting the symbol.</p>
2831
2832</div>
2833
2834<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002835<h3>
2836 <a name="intg_compiler_used">
2837 The '<tt>llvm.compiler.used</tt>' Global Variable
2838 </a>
2839</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +00002840
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002841<div>
Chris Lattner401e10c2009-07-20 06:14:25 +00002842
2843<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2844<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2845touching the symbol. On targets that support it, this allows an intelligent
2846linker to optimize references to the symbol without being impeded as it would be
2847by <tt>@llvm.used</tt>.</p>
2848
2849<p>This is a rare construct that should only be used in rare circumstances, and
2850should not be exposed to source languages.</p>
2851
2852</div>
2853
2854<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002855<h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002856<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002857</h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002858
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002859<div>
David Chisnalle31e9962010-04-30 19:23:49 +00002860<pre>
2861%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +00002862@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +00002863</pre>
2864<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor functions and associated priorities. The functions referenced by this array will be called in ascending order of priority (i.e. lowest first) when the module is loaded. The order of functions with the same priority is not defined.
2865</p>
Chris Lattner857755c2009-07-20 05:55:19 +00002866
2867</div>
2868
2869<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002870<h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002871<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002872</h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002873
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002874<div>
David Chisnalle31e9962010-04-30 19:23:49 +00002875<pre>
2876%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +00002877@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +00002878</pre>
Chris Lattner857755c2009-07-20 05:55:19 +00002879
David Chisnalle31e9962010-04-30 19:23:49 +00002880<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions and associated priorities. The functions referenced by this array will be called in descending order of priority (i.e. highest first) when the module is loaded. The order of functions with the same priority is not defined.
2881</p>
Chris Lattner857755c2009-07-20 05:55:19 +00002882
2883</div>
2884
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002885</div>
Chris Lattner857755c2009-07-20 05:55:19 +00002886
Chris Lattnere87d6532006-01-25 23:47:57 +00002887<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002888<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +00002889<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00002890
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002891<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002892
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002893<p>The LLVM instruction set consists of several different classifications of
2894 instructions: <a href="#terminators">terminator
2895 instructions</a>, <a href="#binaryops">binary instructions</a>,
2896 <a href="#bitwiseops">bitwise binary instructions</a>,
2897 <a href="#memoryops">memory instructions</a>, and
2898 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002899
Chris Lattner00950542001-06-06 20:29:01 +00002900<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002901<h3>
2902 <a name="terminators">Terminator Instructions</a>
2903</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002904
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002905<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002906
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002907<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2908 in a program ends with a "Terminator" instruction, which indicates which
2909 block should be executed after the current block is finished. These
2910 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2911 control flow, not values (the one exception being the
2912 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2913
Duncan Sands83821c82010-04-15 20:35:54 +00002914<p>There are seven different terminator instructions: the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002915 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2916 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2917 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling21c346e2009-11-02 00:25:26 +00002918 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002919 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2920 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2921 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002922
Chris Lattner00950542001-06-06 20:29:01 +00002923<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002924<h4>
2925 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
2926</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002927
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002928<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002929
Chris Lattner00950542001-06-06 20:29:01 +00002930<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +00002931<pre>
2932 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002933 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00002934</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002935
Chris Lattner00950542001-06-06 20:29:01 +00002936<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002937<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2938 a value) from a function back to the caller.</p>
2939
2940<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2941 value and then causes control flow, and one that just causes control flow to
2942 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002943
Chris Lattner00950542001-06-06 20:29:01 +00002944<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002945<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2946 return value. The type of the return value must be a
2947 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +00002948
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002949<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2950 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2951 value or a return value with a type that does not match its type, or if it
2952 has a void return type and contains a '<tt>ret</tt>' instruction with a
2953 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002954
Chris Lattner00950542001-06-06 20:29:01 +00002955<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002956<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2957 the calling function's context. If the caller is a
2958 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2959 instruction after the call. If the caller was an
2960 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2961 the beginning of the "normal" destination block. If the instruction returns
2962 a value, that value shall set the call or invoke instruction's return
2963 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002964
Chris Lattner00950542001-06-06 20:29:01 +00002965<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002966<pre>
2967 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002968 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +00002969 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Chris Lattner00950542001-06-06 20:29:01 +00002970</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +00002971
Misha Brukman9d0919f2003-11-08 01:05:38 +00002972</div>
Chris Lattner00950542001-06-06 20:29:01 +00002973<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002974<h4>
2975 <a name="i_br">'<tt>br</tt>' Instruction</a>
2976</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002977
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002978<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002979
Chris Lattner00950542001-06-06 20:29:01 +00002980<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002981<pre>
2982 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 +00002983</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002984
Chris Lattner00950542001-06-06 20:29:01 +00002985<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002986<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2987 different basic block in the current function. There are two forms of this
2988 instruction, corresponding to a conditional branch and an unconditional
2989 branch.</p>
2990
Chris Lattner00950542001-06-06 20:29:01 +00002991<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002992<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2993 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2994 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2995 target.</p>
2996
Chris Lattner00950542001-06-06 20:29:01 +00002997<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +00002998<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002999 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3000 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3001 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3002
Chris Lattner00950542001-06-06 20:29:01 +00003003<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00003004<pre>
3005Test:
3006 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3007 br i1 %cond, label %IfEqual, label %IfUnequal
3008IfEqual:
3009 <a href="#i_ret">ret</a> i32 1
3010IfUnequal:
3011 <a href="#i_ret">ret</a> i32 0
3012</pre>
3013
Misha Brukman9d0919f2003-11-08 01:05:38 +00003014</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003015
Chris Lattner00950542001-06-06 20:29:01 +00003016<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003017<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003018 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003019</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003020
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003021<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003022
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003023<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003024<pre>
3025 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3026</pre>
3027
Chris Lattner00950542001-06-06 20:29:01 +00003028<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003029<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003030 several different places. It is a generalization of the '<tt>br</tt>'
3031 instruction, allowing a branch to occur to one of many possible
3032 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003033
Chris Lattner00950542001-06-06 20:29:01 +00003034<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003035<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003036 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3037 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3038 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003039
Chris Lattner00950542001-06-06 20:29:01 +00003040<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00003041<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003042 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3043 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +00003044 transferred to the corresponding destination; otherwise, control flow is
3045 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00003046
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003047<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003048<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003049 <tt>switch</tt> instruction, this instruction may be code generated in
3050 different ways. For example, it could be generated as a series of chained
3051 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003052
3053<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003054<pre>
3055 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00003056 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +00003057 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003058
3059 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003060 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003061
3062 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +00003063 switch i32 %val, label %otherwise [ i32 0, label %onzero
3064 i32 1, label %onone
3065 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00003066</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003067
Misha Brukman9d0919f2003-11-08 01:05:38 +00003068</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003069
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003070
3071<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003072<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +00003073 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003074</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003075
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003076<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003077
3078<h5>Syntax:</h5>
3079<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +00003080 indirectbr &lt;somety&gt;* &lt;address&gt;, [ label &lt;dest1&gt;, label &lt;dest2&gt;, ... ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003081</pre>
3082
3083<h5>Overview:</h5>
3084
Chris Lattnerab21db72009-10-28 00:19:10 +00003085<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003086 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +00003087 "<tt>address</tt>". Address must be derived from a <a
3088 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003089
3090<h5>Arguments:</h5>
3091
3092<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3093 rest of the arguments indicate the full set of possible destinations that the
3094 address may point to. Blocks are allowed to occur multiple times in the
3095 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003096
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003097<p>This destination list is required so that dataflow analysis has an accurate
3098 understanding of the CFG.</p>
3099
3100<h5>Semantics:</h5>
3101
3102<p>Control transfers to the block specified in the address argument. All
3103 possible destination blocks must be listed in the label list, otherwise this
3104 instruction has undefined behavior. This implies that jumps to labels
3105 defined in other functions have undefined behavior as well.</p>
3106
3107<h5>Implementation:</h5>
3108
3109<p>This is typically implemented with a jump through a register.</p>
3110
3111<h5>Example:</h5>
3112<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +00003113 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003114</pre>
3115
3116</div>
3117
3118
Chris Lattner00950542001-06-06 20:29:01 +00003119<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003120<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003121 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003122</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003123
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003124<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003125
Chris Lattner00950542001-06-06 20:29:01 +00003126<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003127<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +00003128 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] &lt;ptr to function ty&gt; &lt;function ptr val&gt;(&lt;function args&gt;) [<a href="#fnattrs">fn attrs</a>]
Chris Lattner76b8a332006-05-14 18:23:06 +00003129 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003130</pre>
3131
Chris Lattner6536cfe2002-05-06 22:08:29 +00003132<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003133<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003134 function, with the possibility of control flow transfer to either the
3135 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3136 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3137 control flow will return to the "normal" label. If the callee (or any
3138 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3139 instruction, control is interrupted and continued at the dynamically nearest
3140 "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003141
Chris Lattner00950542001-06-06 20:29:01 +00003142<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003143<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003144
Chris Lattner00950542001-06-06 20:29:01 +00003145<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003146 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3147 convention</a> the call should use. If none is specified, the call
3148 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +00003149
3150 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003151 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3152 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +00003153
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003154 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003155 function value being invoked. In most cases, this is a direct function
3156 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3157 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003158
3159 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003160 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003161
3162 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +00003163 signature argument types and parameter attributes. All arguments must be
3164 of <a href="#t_firstclass">first class</a> type. If the function
3165 signature indicates the function accepts a variable number of arguments,
3166 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003167
3168 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003169 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003170
3171 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003172 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003173
Devang Patel307e8ab2008-10-07 17:48:33 +00003174 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003175 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3176 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +00003177</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003178
Chris Lattner00950542001-06-06 20:29:01 +00003179<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003180<p>This instruction is designed to operate as a standard
3181 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3182 primary difference is that it establishes an association with a label, which
3183 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003184
3185<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003186 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3187 exception. Additionally, this is important for implementation of
3188 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003189
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003190<p>For the purposes of the SSA form, the definition of the value returned by the
3191 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3192 block to the "normal" label. If the callee unwinds then no return value is
3193 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +00003194
Chris Lattnerdf7a6802010-01-15 18:08:37 +00003195<p>Note that the code generator does not yet completely support unwind, and
3196that the invoke/unwind semantics are likely to change in future versions.</p>
3197
Chris Lattner00950542001-06-06 20:29:01 +00003198<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003199<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +00003200 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003201 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +00003202 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003203 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00003204</pre>
Chris Lattner35eca582004-10-16 18:04:13 +00003205
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003206</div>
Chris Lattner35eca582004-10-16 18:04:13 +00003207
Chris Lattner27f71f22003-09-03 00:41:47 +00003208<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00003209
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003210<h4>
3211 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3212</h4>
Chris Lattner35eca582004-10-16 18:04:13 +00003213
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003214<div>
Chris Lattner35eca582004-10-16 18:04:13 +00003215
Chris Lattner27f71f22003-09-03 00:41:47 +00003216<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003217<pre>
3218 unwind
3219</pre>
3220
Chris Lattner27f71f22003-09-03 00:41:47 +00003221<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003222<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003223 at the first callee in the dynamic call stack which used
3224 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3225 This is primarily used to implement exception handling.</p>
Chris Lattner35eca582004-10-16 18:04:13 +00003226
Chris Lattner27f71f22003-09-03 00:41:47 +00003227<h5>Semantics:</h5>
Chris Lattner72ed2002008-04-19 21:01:16 +00003228<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003229 immediately halt. The dynamic call stack is then searched for the
3230 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3231 Once found, execution continues at the "exceptional" destination block
3232 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3233 instruction in the dynamic call chain, undefined behavior results.</p>
3234
Chris Lattnerdf7a6802010-01-15 18:08:37 +00003235<p>Note that the code generator does not yet completely support unwind, and
3236that the invoke/unwind semantics are likely to change in future versions.</p>
3237
Misha Brukman9d0919f2003-11-08 01:05:38 +00003238</div>
Chris Lattner35eca582004-10-16 18:04:13 +00003239
3240<!-- _______________________________________________________________________ -->
3241
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003242<h4>
3243 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3244</h4>
Chris Lattner35eca582004-10-16 18:04:13 +00003245
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003246<div>
Chris Lattner35eca582004-10-16 18:04:13 +00003247
3248<h5>Syntax:</h5>
3249<pre>
3250 unreachable
3251</pre>
3252
3253<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003254<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003255 instruction is used to inform the optimizer that a particular portion of the
3256 code is not reachable. This can be used to indicate that the code after a
3257 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +00003258
3259<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003260<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003261
Chris Lattner35eca582004-10-16 18:04:13 +00003262</div>
3263
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003264</div>
3265
Chris Lattner00950542001-06-06 20:29:01 +00003266<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003267<h3>
3268 <a name="binaryops">Binary Operations</a>
3269</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003270
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003271<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003272
3273<p>Binary operators are used to do most of the computation in a program. They
3274 require two operands of the same type, execute an operation on them, and
3275 produce a single value. The operands might represent multiple data, as is
3276 the case with the <a href="#t_vector">vector</a> data type. The result value
3277 has the same type as its operands.</p>
3278
Misha Brukman9d0919f2003-11-08 01:05:38 +00003279<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003280
Chris Lattner00950542001-06-06 20:29:01 +00003281<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003282<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003283 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003284</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003285
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003286<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003287
Chris Lattner00950542001-06-06 20:29:01 +00003288<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003289<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003290 &lt;result&gt; = add &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanfdfca792009-09-02 17:31:42 +00003291 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3292 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3293 &lt;result&gt; = add nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00003294</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003295
Chris Lattner00950542001-06-06 20:29:01 +00003296<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003297<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003298
Chris Lattner00950542001-06-06 20:29:01 +00003299<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003300<p>The two arguments to the '<tt>add</tt>' instruction must
3301 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3302 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003303
Chris Lattner00950542001-06-06 20:29:01 +00003304<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003305<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003306
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003307<p>If the sum has unsigned overflow, the result returned is the mathematical
3308 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003309
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003310<p>Because LLVM integers use a two's complement representation, this instruction
3311 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003312
Dan Gohman08d012e2009-07-22 22:44:56 +00003313<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3314 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3315 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +00003316 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3317 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003318
Chris Lattner00950542001-06-06 20:29:01 +00003319<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003320<pre>
3321 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00003322</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003323
Misha Brukman9d0919f2003-11-08 01:05:38 +00003324</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003325
Chris Lattner00950542001-06-06 20:29:01 +00003326<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003327<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003328 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003329</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003330
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003331<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003332
3333<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003334<pre>
3335 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3336</pre>
3337
3338<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003339<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3340
3341<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003342<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003343 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3344 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003345
3346<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003347<p>The value produced is the floating point sum of the two operands.</p>
3348
3349<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003350<pre>
3351 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3352</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003353
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003354</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003355
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003356<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003357<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003358 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003359</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003360
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003361<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003362
Chris Lattner00950542001-06-06 20:29:01 +00003363<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003364<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +00003365 &lt;result&gt; = sub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanfdfca792009-09-02 17:31:42 +00003366 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3367 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3368 &lt;result&gt; = sub nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00003369</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003370
Chris Lattner00950542001-06-06 20:29:01 +00003371<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003372<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003373 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003374
3375<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003376 '<tt>neg</tt>' instruction present in most other intermediate
3377 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003378
Chris Lattner00950542001-06-06 20:29:01 +00003379<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003380<p>The two arguments to the '<tt>sub</tt>' instruction must
3381 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3382 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003383
Chris Lattner00950542001-06-06 20:29:01 +00003384<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003385<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003386
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003387<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003388 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3389 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003390
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003391<p>Because LLVM integers use a two's complement representation, this instruction
3392 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003393
Dan Gohman08d012e2009-07-22 22:44:56 +00003394<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3395 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3396 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +00003397 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3398 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003399
Chris Lattner00950542001-06-06 20:29:01 +00003400<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +00003401<pre>
3402 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003403 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +00003404</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003405
Misha Brukman9d0919f2003-11-08 01:05:38 +00003406</div>
Chris Lattner5568e942008-05-20 20:48:21 +00003407
Chris Lattner00950542001-06-06 20:29:01 +00003408<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003409<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003410 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003411</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003412
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003413<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003414
3415<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003416<pre>
3417 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3418</pre>
3419
3420<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003421<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003422 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003423
3424<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003425 '<tt>fneg</tt>' instruction present in most other intermediate
3426 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003427
3428<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +00003429<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003430 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3431 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003432
3433<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003434<p>The value produced is the floating point difference of the two operands.</p>
3435
3436<h5>Example:</h5>
3437<pre>
3438 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3439 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3440</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003441
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003442</div>
3443
3444<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003445<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003446 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003447</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003448
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003449<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003450
Chris Lattner00950542001-06-06 20:29:01 +00003451<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003452<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +00003453 &lt;result&gt; = mul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanfdfca792009-09-02 17:31:42 +00003454 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3455 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3456 &lt;result&gt; = mul nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00003457</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003458
Chris Lattner00950542001-06-06 20:29:01 +00003459<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003460<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003461
Chris Lattner00950542001-06-06 20:29:01 +00003462<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003463<p>The two arguments to the '<tt>mul</tt>' instruction must
3464 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3465 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003466
Chris Lattner00950542001-06-06 20:29:01 +00003467<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003468<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003469
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003470<p>If the result of the multiplication has unsigned overflow, the result
3471 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3472 width of the result.</p>
3473
3474<p>Because LLVM integers use a two's complement representation, and the result
3475 is the same width as the operands, this instruction returns the correct
3476 result for both signed and unsigned integers. If a full product
3477 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3478 be sign-extended or zero-extended as appropriate to the width of the full
3479 product.</p>
3480
Dan Gohman08d012e2009-07-22 22:44:56 +00003481<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3482 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3483 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +00003484 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3485 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003486
Chris Lattner00950542001-06-06 20:29:01 +00003487<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003488<pre>
3489 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00003490</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003491
Misha Brukman9d0919f2003-11-08 01:05:38 +00003492</div>
Chris Lattner5568e942008-05-20 20:48:21 +00003493
Chris Lattner00950542001-06-06 20:29:01 +00003494<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003495<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003496 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003497</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003498
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003499<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003500
3501<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003502<pre>
3503 &lt;result&gt; = fmul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003504</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003505
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003506<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003507<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003508
3509<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003510<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003511 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3512 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003513
3514<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003515<p>The value produced is the floating point product of the two operands.</p>
3516
3517<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003518<pre>
3519 &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003520</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003521
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003522</div>
3523
3524<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003525<h4>
3526 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3527</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003528
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003529<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003530
Reid Spencer1628cec2006-10-26 06:15:43 +00003531<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003532<pre>
Chris Lattner35bda892011-02-06 21:44:57 +00003533 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3534 &lt;result&gt; = udiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00003535</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003536
Reid Spencer1628cec2006-10-26 06:15:43 +00003537<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003538<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003539
Reid Spencer1628cec2006-10-26 06:15:43 +00003540<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003541<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003542 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3543 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003544
Reid Spencer1628cec2006-10-26 06:15:43 +00003545<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +00003546<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003547
Chris Lattner5ec89832008-01-28 00:36:27 +00003548<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003549 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3550
Chris Lattner5ec89832008-01-28 00:36:27 +00003551<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003552
Chris Lattner35bda892011-02-06 21:44:57 +00003553<p>If the <tt>exact</tt> keyword is present, the result value of the
3554 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3555 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3556
3557
Reid Spencer1628cec2006-10-26 06:15:43 +00003558<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003559<pre>
3560 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00003561</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003562
Reid Spencer1628cec2006-10-26 06:15:43 +00003563</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003564
Reid Spencer1628cec2006-10-26 06:15:43 +00003565<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003566<h4>
3567 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3568</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003569
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003570<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003571
Reid Spencer1628cec2006-10-26 06:15:43 +00003572<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003573<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003574 &lt;result&gt; = sdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Dan Gohmanfdfca792009-09-02 17:31:42 +00003575 &lt;result&gt; = sdiv exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00003576</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003577
Reid Spencer1628cec2006-10-26 06:15:43 +00003578<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003579<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003580
Reid Spencer1628cec2006-10-26 06:15:43 +00003581<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003582<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003583 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3584 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003585
Reid Spencer1628cec2006-10-26 06:15:43 +00003586<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003587<p>The value produced is the signed integer quotient of the two operands rounded
3588 towards zero.</p>
3589
Chris Lattner5ec89832008-01-28 00:36:27 +00003590<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003591 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3592
Chris Lattner5ec89832008-01-28 00:36:27 +00003593<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003594 undefined behavior; this is a rare case, but can occur, for example, by doing
3595 a 32-bit division of -2147483648 by -1.</p>
3596
Dan Gohman9c5beed2009-07-22 00:04:19 +00003597<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +00003598 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohman38da9272010-07-11 00:08:34 +00003599 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003600
Reid Spencer1628cec2006-10-26 06:15:43 +00003601<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003602<pre>
3603 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00003604</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003605
Reid Spencer1628cec2006-10-26 06:15:43 +00003606</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003607
Reid Spencer1628cec2006-10-26 06:15:43 +00003608<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003609<h4>
3610 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3611</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003612
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003613<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003614
Chris Lattner00950542001-06-06 20:29:01 +00003615<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003616<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +00003617 &lt;result&gt; = fdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +00003618</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003619
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003620<h5>Overview:</h5>
3621<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003622
Chris Lattner261efe92003-11-25 01:02:51 +00003623<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003624<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003625 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3626 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003627
Chris Lattner261efe92003-11-25 01:02:51 +00003628<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00003629<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003630
Chris Lattner261efe92003-11-25 01:02:51 +00003631<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003632<pre>
3633 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00003634</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003635
Chris Lattner261efe92003-11-25 01:02:51 +00003636</div>
Chris Lattner5568e942008-05-20 20:48:21 +00003637
Chris Lattner261efe92003-11-25 01:02:51 +00003638<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003639<h4>
3640 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3641</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003642
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003643<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003644
Reid Spencer0a783f72006-11-02 01:53:59 +00003645<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003646<pre>
3647 &lt;result&gt; = urem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00003648</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003649
Reid Spencer0a783f72006-11-02 01:53:59 +00003650<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003651<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3652 division of its two arguments.</p>
3653
Reid Spencer0a783f72006-11-02 01:53:59 +00003654<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003655<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003656 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3657 values. Both arguments must have identical types.</p>
3658
Reid Spencer0a783f72006-11-02 01:53:59 +00003659<h5>Semantics:</h5>
3660<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003661 This instruction always performs an unsigned division to get the
3662 remainder.</p>
3663
Chris Lattner5ec89832008-01-28 00:36:27 +00003664<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003665 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3666
Chris Lattner5ec89832008-01-28 00:36:27 +00003667<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003668
Reid Spencer0a783f72006-11-02 01:53:59 +00003669<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003670<pre>
3671 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00003672</pre>
3673
3674</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003675
Reid Spencer0a783f72006-11-02 01:53:59 +00003676<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003677<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003678 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003679</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003680
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003681<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003682
Chris Lattner261efe92003-11-25 01:02:51 +00003683<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003684<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +00003685 &lt;result&gt; = srem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +00003686</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003687
Chris Lattner261efe92003-11-25 01:02:51 +00003688<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003689<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3690 division of its two operands. This instruction can also take
3691 <a href="#t_vector">vector</a> versions of the values in which case the
3692 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +00003693
Chris Lattner261efe92003-11-25 01:02:51 +00003694<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003695<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003696 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3697 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003698
Chris Lattner261efe92003-11-25 01:02:51 +00003699<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00003700<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +00003701 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3702 <i>modulo</i> operator (where the result is either zero or has the same sign
3703 as the divisor, <tt>op2</tt>) of a value.
3704 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003705 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3706 Math Forum</a>. For a table of how this is implemented in various languages,
3707 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3708 Wikipedia: modulo operation</a>.</p>
3709
Chris Lattner5ec89832008-01-28 00:36:27 +00003710<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003711 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3712
Chris Lattner5ec89832008-01-28 00:36:27 +00003713<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003714 Overflow also leads to undefined behavior; this is a rare case, but can
3715 occur, for example, by taking the remainder of a 32-bit division of
3716 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3717 lets srem be implemented using instructions that return both the result of
3718 the division and the remainder.)</p>
3719
Chris Lattner261efe92003-11-25 01:02:51 +00003720<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003721<pre>
3722 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00003723</pre>
3724
3725</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003726
Reid Spencer0a783f72006-11-02 01:53:59 +00003727<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003728<h4>
3729 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3730</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003731
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003732<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003733
Reid Spencer0a783f72006-11-02 01:53:59 +00003734<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003735<pre>
3736 &lt;result&gt; = frem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00003737</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003738
Reid Spencer0a783f72006-11-02 01:53:59 +00003739<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003740<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3741 its two operands.</p>
3742
Reid Spencer0a783f72006-11-02 01:53:59 +00003743<h5>Arguments:</h5>
3744<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003745 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3746 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003747
Reid Spencer0a783f72006-11-02 01:53:59 +00003748<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003749<p>This instruction returns the <i>remainder</i> of a division. The remainder
3750 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003751
Reid Spencer0a783f72006-11-02 01:53:59 +00003752<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003753<pre>
3754 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00003755</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003756
Misha Brukman9d0919f2003-11-08 01:05:38 +00003757</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +00003758
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003759</div>
3760
Reid Spencer8e11bf82007-02-02 13:57:07 +00003761<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003762<h3>
3763 <a name="bitwiseops">Bitwise Binary Operations</a>
3764</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003765
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003766<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003767
3768<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3769 program. They are generally very efficient instructions and can commonly be
3770 strength reduced from other instructions. They require two operands of the
3771 same type, execute an operation on them, and produce a single value. The
3772 resulting value is the same type as its operands.</p>
3773
Reid Spencer569f2fa2007-01-31 21:39:12 +00003774<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003775<h4>
3776 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3777</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003778
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003779<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003780
Reid Spencer569f2fa2007-01-31 21:39:12 +00003781<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003782<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +00003783 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3784 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3785 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3786 &lt;result&gt; = shl nuw nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003787</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003788
Reid Spencer569f2fa2007-01-31 21:39:12 +00003789<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003790<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3791 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003792
Reid Spencer569f2fa2007-01-31 21:39:12 +00003793<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003794<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3795 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3796 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003797
Reid Spencer569f2fa2007-01-31 21:39:12 +00003798<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003799<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3800 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3801 is (statically or dynamically) negative or equal to or larger than the number
3802 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3803 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3804 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003805
Chris Lattnerf067d582011-02-07 16:40:21 +00003806<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3807 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
Chris Lattner66298c12011-02-09 16:44:44 +00003808 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnerf067d582011-02-07 16:40:21 +00003809 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3810 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3811 they would if the shift were expressed as a mul instruction with the same
3812 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3813
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003814<h5>Example:</h5>
3815<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003816 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3817 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3818 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003819 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +00003820 &lt;result&gt; = shl &lt;2 x i32&gt; &lt; i32 1, i32 1&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 2, i32 4&gt;</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003821</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003822
Reid Spencer569f2fa2007-01-31 21:39:12 +00003823</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003824
Reid Spencer569f2fa2007-01-31 21:39:12 +00003825<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003826<h4>
3827 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
3828</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003829
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003830<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003831
Reid Spencer569f2fa2007-01-31 21:39:12 +00003832<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003833<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +00003834 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3835 &lt;result&gt; = lshr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003836</pre>
3837
3838<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003839<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3840 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003841
3842<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003843<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003844 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3845 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003846
3847<h5>Semantics:</h5>
3848<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003849 significant bits of the result will be filled with zero bits after the shift.
3850 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3851 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3852 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3853 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003854
Chris Lattnerf067d582011-02-07 16:40:21 +00003855<p>If the <tt>exact</tt> keyword is present, the result value of the
3856 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3857 shifted out are non-zero.</p>
3858
3859
Reid Spencer569f2fa2007-01-31 21:39:12 +00003860<h5>Example:</h5>
3861<pre>
3862 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3863 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3864 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3865 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003866 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +00003867 &lt;result&gt; = lshr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0x7FFFFFFF, i32 1&gt;</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003868</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003869
Reid Spencer569f2fa2007-01-31 21:39:12 +00003870</div>
3871
Reid Spencer8e11bf82007-02-02 13:57:07 +00003872<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003873<h4>
3874 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
3875</h4>
3876
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003877<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003878
3879<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003880<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +00003881 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3882 &lt;result&gt; = ashr exact &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003883</pre>
3884
3885<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003886<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3887 operand shifted to the right a specified number of bits with sign
3888 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003889
3890<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003891<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003892 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3893 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003894
3895<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003896<p>This instruction always performs an arithmetic shift right operation, The
3897 most significant bits of the result will be filled with the sign bit
3898 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3899 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3900 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3901 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003902
Chris Lattnerf067d582011-02-07 16:40:21 +00003903<p>If the <tt>exact</tt> keyword is present, the result value of the
3904 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3905 shifted out are non-zero.</p>
3906
Reid Spencer569f2fa2007-01-31 21:39:12 +00003907<h5>Example:</h5>
3908<pre>
3909 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3910 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3911 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3912 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003913 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +00003914 &lt;result&gt; = ashr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 3&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 -1, i32 0&gt;</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003915</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003916
Reid Spencer569f2fa2007-01-31 21:39:12 +00003917</div>
3918
Chris Lattner00950542001-06-06 20:29:01 +00003919<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003920<h4>
3921 <a name="i_and">'<tt>and</tt>' Instruction</a>
3922</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003923
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003924<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003925
Chris Lattner00950542001-06-06 20:29:01 +00003926<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003927<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +00003928 &lt;result&gt; = and &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00003929</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003930
Chris Lattner00950542001-06-06 20:29:01 +00003931<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003932<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3933 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003934
Chris Lattner00950542001-06-06 20:29:01 +00003935<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003936<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003937 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3938 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003939
Chris Lattner00950542001-06-06 20:29:01 +00003940<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003941<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003942
Misha Brukman9d0919f2003-11-08 01:05:38 +00003943<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00003944 <tbody>
3945 <tr>
3946 <td>In0</td>
3947 <td>In1</td>
3948 <td>Out</td>
3949 </tr>
3950 <tr>
3951 <td>0</td>
3952 <td>0</td>
3953 <td>0</td>
3954 </tr>
3955 <tr>
3956 <td>0</td>
3957 <td>1</td>
3958 <td>0</td>
3959 </tr>
3960 <tr>
3961 <td>1</td>
3962 <td>0</td>
3963 <td>0</td>
3964 </tr>
3965 <tr>
3966 <td>1</td>
3967 <td>1</td>
3968 <td>1</td>
3969 </tr>
3970 </tbody>
3971</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003972
Chris Lattner00950542001-06-06 20:29:01 +00003973<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003974<pre>
3975 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003976 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3977 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +00003978</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003979</div>
Chris Lattner00950542001-06-06 20:29:01 +00003980<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003981<h4>
3982 <a name="i_or">'<tt>or</tt>' Instruction</a>
3983</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003984
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003985<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003986
3987<h5>Syntax:</h5>
3988<pre>
3989 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3990</pre>
3991
3992<h5>Overview:</h5>
3993<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3994 two operands.</p>
3995
3996<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003997<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003998 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3999 values. Both arguments must have identical types.</p>
4000
Chris Lattner00950542001-06-06 20:29:01 +00004001<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004002<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004003
Chris Lattner261efe92003-11-25 01:02:51 +00004004<table border="1" cellspacing="0" cellpadding="4">
4005 <tbody>
4006 <tr>
4007 <td>In0</td>
4008 <td>In1</td>
4009 <td>Out</td>
4010 </tr>
4011 <tr>
4012 <td>0</td>
4013 <td>0</td>
4014 <td>0</td>
4015 </tr>
4016 <tr>
4017 <td>0</td>
4018 <td>1</td>
4019 <td>1</td>
4020 </tr>
4021 <tr>
4022 <td>1</td>
4023 <td>0</td>
4024 <td>1</td>
4025 </tr>
4026 <tr>
4027 <td>1</td>
4028 <td>1</td>
4029 <td>1</td>
4030 </tr>
4031 </tbody>
4032</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004033
Chris Lattner00950542001-06-06 20:29:01 +00004034<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004035<pre>
4036 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004037 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4038 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +00004039</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004040
Misha Brukman9d0919f2003-11-08 01:05:38 +00004041</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004042
Chris Lattner00950542001-06-06 20:29:01 +00004043<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004044<h4>
4045 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4046</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004047
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004048<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004049
Chris Lattner00950542001-06-06 20:29:01 +00004050<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004051<pre>
4052 &lt;result&gt; = xor &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00004053</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004054
Chris Lattner00950542001-06-06 20:29:01 +00004055<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004056<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4057 its two operands. The <tt>xor</tt> is used to implement the "one's
4058 complement" operation, which is the "~" operator in C.</p>
4059
Chris Lattner00950542001-06-06 20:29:01 +00004060<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004061<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004062 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4063 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00004064
Chris Lattner00950542001-06-06 20:29:01 +00004065<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004066<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004067
Chris Lattner261efe92003-11-25 01:02:51 +00004068<table border="1" cellspacing="0" cellpadding="4">
4069 <tbody>
4070 <tr>
4071 <td>In0</td>
4072 <td>In1</td>
4073 <td>Out</td>
4074 </tr>
4075 <tr>
4076 <td>0</td>
4077 <td>0</td>
4078 <td>0</td>
4079 </tr>
4080 <tr>
4081 <td>0</td>
4082 <td>1</td>
4083 <td>1</td>
4084 </tr>
4085 <tr>
4086 <td>1</td>
4087 <td>0</td>
4088 <td>1</td>
4089 </tr>
4090 <tr>
4091 <td>1</td>
4092 <td>1</td>
4093 <td>0</td>
4094 </tr>
4095 </tbody>
4096</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004097
Chris Lattner00950542001-06-06 20:29:01 +00004098<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004099<pre>
4100 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004101 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4102 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4103 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00004104</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004105
Misha Brukman9d0919f2003-11-08 01:05:38 +00004106</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004107
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004108</div>
4109
Chris Lattner00950542001-06-06 20:29:01 +00004110<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004111<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +00004112 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004113</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +00004114
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004115<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004116
4117<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004118 target-independent manner. These instructions cover the element-access and
4119 vector-specific operations needed to process vectors effectively. While LLVM
4120 does directly support these vector operations, many sophisticated algorithms
4121 will want to use target-specific intrinsics to take full advantage of a
4122 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004123
Chris Lattner3df241e2006-04-08 23:07:04 +00004124<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004125<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004126 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004127</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004128
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004129<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004130
4131<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004132<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004133 &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 +00004134</pre>
4135
4136<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004137<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4138 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004139
4140
4141<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004142<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4143 of <a href="#t_vector">vector</a> type. The second operand is an index
4144 indicating the position from which to extract the element. The index may be
4145 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004146
4147<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004148<p>The result is a scalar of the same type as the element type of
4149 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4150 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4151 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004152
4153<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004154<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +00004155 &lt;result&gt; = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00004156</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +00004157
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004158</div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004159
4160<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004161<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004162 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004163</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004164
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004165<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004166
4167<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004168<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +00004169 &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 +00004170</pre>
4171
4172<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004173<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4174 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004175
4176<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004177<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4178 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4179 whose type must equal the element type of the first operand. The third
4180 operand is an index indicating the position at which to insert the value.
4181 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004182
4183<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004184<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4185 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4186 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4187 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004188
4189<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004190<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +00004191 &lt;result&gt; = 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 +00004192</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004193
Chris Lattner3df241e2006-04-08 23:07:04 +00004194</div>
4195
4196<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004197<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004198 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004199</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004200
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004201<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004202
4203<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004204<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +00004205 &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;m x i32&gt; &lt;mask&gt; <i>; yields &lt;m x &lt;ty&gt;&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00004206</pre>
4207
4208<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004209<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4210 from two input vectors, returning a vector with the same element type as the
4211 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004212
4213<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004214<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4215 with types that match each other. The third argument is a shuffle mask whose
4216 element type is always 'i32'. The result of the instruction is a vector
4217 whose length is the same as the shuffle mask and whose element type is the
4218 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004219
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004220<p>The shuffle mask operand is required to be a constant vector with either
4221 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004222
4223<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004224<p>The elements of the two input vectors are numbered from left to right across
4225 both of the vectors. The shuffle mask operand specifies, for each element of
4226 the result vector, which element of the two input vectors the result element
4227 gets. The element selector may be undef (meaning "don't care") and the
4228 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004229
4230<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004231<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004232 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen6f1cc772007-04-22 01:17:39 +00004233 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004234 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
Reid Spencerca86e162006-12-31 07:07:53 +00004235 &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.
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004236 &lt;result&gt; = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
Mon P Wangaeb06d22008-11-10 04:46:22 +00004237 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004238 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Mon P Wangaeb06d22008-11-10 04:46:22 +00004239 &lt;8 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 &gt; <i>; yields &lt;8 x i32&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00004240</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +00004241
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004242</div>
Tanya Lattner09474292006-04-14 19:24:33 +00004243
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004244</div>
4245
Chris Lattner3df241e2006-04-08 23:07:04 +00004246<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004247<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004248 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004249</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004250
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004251<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004252
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004253<p>LLVM supports several instructions for working with
4254 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004255
Dan Gohmana334d5f2008-05-12 23:51:09 +00004256<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004257<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004258 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004259</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004260
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004261<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004262
4263<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004264<pre>
4265 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4266</pre>
4267
4268<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004269<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4270 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004271
4272<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004273<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +00004274 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004275 <a href="#t_array">array</a> type. The operands are constant indices to
4276 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004277 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +00004278 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4279 <ul>
4280 <li>Since the value being indexed is not a pointer, the first index is
4281 omitted and assumed to be zero.</li>
4282 <li>At least one index must be specified.</li>
4283 <li>Not only struct indices but also array indices must be in
4284 bounds.</li>
4285 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004286
4287<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004288<p>The result is the value at the position in the aggregate specified by the
4289 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004290
4291<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004292<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +00004293 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004294</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004295
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004296</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004297
4298<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004299<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004300 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004301</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004302
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004303<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004304
4305<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004306<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +00004307 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt;{, <idx>}* <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004308</pre>
4309
4310<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004311<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4312 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004313
4314<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004315<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +00004316 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004317 <a href="#t_array">array</a> type. The second operand is a first-class
4318 value to insert. The following operands are constant indices indicating
4319 the position at which to insert the value in a similar manner as indices in a
Frits van Bommel13242892010-12-05 20:54:38 +00004320 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004321 value to insert must have the same type as the value identified by the
4322 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004323
4324<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004325<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4326 that of <tt>val</tt> except that the value at the position specified by the
4327 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004328
4329<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004330<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +00004331 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4332 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4333 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 <i>; yields {i32 1, float %val}</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004334</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004335
Dan Gohmana334d5f2008-05-12 23:51:09 +00004336</div>
4337
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004338</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004339
4340<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004341<h3>
Chris Lattner884a9702006-08-15 00:45:58 +00004342 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004343</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004344
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004345<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004346
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004347<p>A key design point of an SSA-based representation is how it represents
4348 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +00004349 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004350 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004351
Chris Lattner00950542001-06-06 20:29:01 +00004352<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004353<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004354 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004355</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004356
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004357<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004358
Chris Lattner00950542001-06-06 20:29:01 +00004359<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004360<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +00004361 &lt;result&gt; = alloca &lt;type&gt;[, &lt;ty&gt; &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +00004362</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004363
Chris Lattner00950542001-06-06 20:29:01 +00004364<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00004365<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004366 currently executing function, to be automatically released when this function
4367 returns to its caller. The object is always allocated in the generic address
4368 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004369
Chris Lattner00950542001-06-06 20:29:01 +00004370<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004371<p>The '<tt>alloca</tt>' instruction
4372 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4373 runtime stack, returning a pointer of the appropriate type to the program.
4374 If "NumElements" is specified, it is the number of elements allocated,
4375 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4376 specified, the value result of the allocation is guaranteed to be aligned to
4377 at least that boundary. If not specified, or if zero, the target can choose
4378 to align the allocation on any convenient boundary compatible with the
4379 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004380
Misha Brukman9d0919f2003-11-08 01:05:38 +00004381<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004382
Chris Lattner00950542001-06-06 20:29:01 +00004383<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +00004384<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004385 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4386 memory is automatically released when the function returns. The
4387 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4388 variables that must have an address available. When the function returns
4389 (either with the <tt><a href="#i_ret">ret</a></tt>
4390 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4391 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004392
Chris Lattner00950542001-06-06 20:29:01 +00004393<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004394<pre>
Dan Gohman81e21672009-01-04 23:49:44 +00004395 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4396 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4397 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4398 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00004399</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004400
Misha Brukman9d0919f2003-11-08 01:05:38 +00004401</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004402
Chris Lattner00950542001-06-06 20:29:01 +00004403<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004404<h4>
4405 <a name="i_load">'<tt>load</tt>' Instruction</a>
4406</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004407
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004408<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004409
Chris Lattner2b7d3202002-05-06 03:03:22 +00004410<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004411<pre>
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004412 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4413 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4414 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004415</pre>
4416
Chris Lattner2b7d3202002-05-06 03:03:22 +00004417<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004418<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004419
Chris Lattner2b7d3202002-05-06 03:03:22 +00004420<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004421<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4422 from which to load. The pointer must point to
4423 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4424 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00004425 number or order of execution of this <tt>load</tt> with other <a
4426 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004427
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004428<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004429 operation (that is, the alignment of the memory address). A value of 0 or an
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004430 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004431 alignment for the target. It is the responsibility of the code emitter to
4432 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004433 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004434 produce less efficient code. An alignment of 1 is always safe.</p>
4435
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004436<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4437 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004438 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004439 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4440 and code generator that this load is not expected to be reused in the cache.
4441 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004442 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +00004443
Chris Lattner2b7d3202002-05-06 03:03:22 +00004444<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004445<p>The location of memory pointed to is loaded. If the value being loaded is of
4446 scalar type then the number of bytes read does not exceed the minimum number
4447 of bytes needed to hold all bits of the type. For example, loading an
4448 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4449 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4450 is undefined if the value was not originally written using a store of the
4451 same type.</p>
4452
Chris Lattner2b7d3202002-05-06 03:03:22 +00004453<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004454<pre>
4455 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4456 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004457 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00004458</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004459
Misha Brukman9d0919f2003-11-08 01:05:38 +00004460</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004461
Chris Lattner2b7d3202002-05-06 03:03:22 +00004462<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004463<h4>
4464 <a name="i_store">'<tt>store</tt>' Instruction</a>
4465</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004466
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004467<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004468
Chris Lattner2b7d3202002-05-06 03:03:22 +00004469<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004470<pre>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004471 store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
4472 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00004473</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004474
Chris Lattner2b7d3202002-05-06 03:03:22 +00004475<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004476<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004477
Chris Lattner2b7d3202002-05-06 03:03:22 +00004478<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004479<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4480 and an address at which to store it. The type of the
4481 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4482 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00004483 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4484 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4485 order of execution of this <tt>store</tt> with other <a
4486 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004487
4488<p>The optional constant "align" argument specifies the alignment of the
4489 operation (that is, the alignment of the memory address). A value of 0 or an
4490 omitted "align" argument means that the operation has the preferential
4491 alignment for the target. It is the responsibility of the code emitter to
4492 ensure that the alignment information is correct. Overestimating the
4493 alignment results in an undefined behavior. Underestimating the alignment may
4494 produce less efficient code. An alignment of 1 is always safe.</p>
4495
David Greene8939b0d2010-02-16 20:50:18 +00004496<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004497 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004498 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +00004499 instruction tells the optimizer and code generator that this load is
4500 not expected to be reused in the cache. The code generator may
4501 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004502 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +00004503
4504
Chris Lattner261efe92003-11-25 01:02:51 +00004505<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004506<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4507 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4508 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4509 does not exceed the minimum number of bytes needed to hold all bits of the
4510 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4511 writing a value of a type like <tt>i20</tt> with a size that is not an
4512 integral number of bytes, it is unspecified what happens to the extra bits
4513 that do not belong to the type, but they will typically be overwritten.</p>
4514
Chris Lattner2b7d3202002-05-06 03:03:22 +00004515<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004516<pre>
4517 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +00004518 store i32 3, i32* %ptr <i>; yields {void}</i>
4519 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00004520</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004521
Reid Spencer47ce1792006-11-09 21:15:49 +00004522</div>
4523
Chris Lattner2b7d3202002-05-06 03:03:22 +00004524<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004525<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004526 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004527</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004528
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004529<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004530
Chris Lattner7faa8832002-04-14 06:13:44 +00004531<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004532<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004533 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohmandd8004d2009-07-27 21:53:46 +00004534 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004535</pre>
4536
Chris Lattner7faa8832002-04-14 06:13:44 +00004537<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004538<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004539 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4540 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004541
Chris Lattner7faa8832002-04-14 06:13:44 +00004542<h5>Arguments:</h5>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004543<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +00004544 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004545 elements of the aggregate object are indexed. The interpretation of each
4546 index is dependent on the type being indexed into. The first index always
4547 indexes the pointer value given as the first argument, the second index
4548 indexes a value of the type pointed to (not necessarily the value directly
4549 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004550 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +00004551 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004552 can never be pointers, since that would require loading the pointer before
4553 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004554
4555<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner61c70e92010-08-28 04:09:24 +00004556 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004557 integer <b>constants</b> are allowed. When indexing into an array, pointer
4558 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnerc8eef442009-07-29 06:44:13 +00004559 constant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004560
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004561<p>For example, let's consider a C code fragment and how it gets compiled to
4562 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004563
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004564<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004565struct RT {
4566 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +00004567 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004568 char C;
4569};
4570struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +00004571 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004572 double Y;
4573 struct RT Z;
4574};
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004575
Chris Lattnercabc8462007-05-29 15:43:56 +00004576int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004577 return &amp;s[1].Z.B[5][13];
4578}
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004579</pre>
4580
Misha Brukman9d0919f2003-11-08 01:05:38 +00004581<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004582
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004583<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +00004584%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4585%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004586
Dan Gohman4df605b2009-07-25 02:23:48 +00004587define i32* @foo(%ST* %s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004588entry:
4589 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4590 ret i32* %reg
4591}
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004592</pre>
4593
Chris Lattner7faa8832002-04-14 06:13:44 +00004594<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004595<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004596 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4597 }</tt>' type, a structure. The second index indexes into the third element
4598 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4599 i8 }</tt>' type, another structure. The third index indexes into the second
4600 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4601 array. The two dimensions of the array are subscripted into, yielding an
4602 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4603 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004604
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004605<p>Note that it is perfectly legal to index partially through a structure,
4606 returning a pointer to an inner element. Because of this, the LLVM code for
4607 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004608
4609<pre>
Dan Gohman4df605b2009-07-25 02:23:48 +00004610 define i32* @foo(%ST* %s) {
Reid Spencerca86e162006-12-31 07:07:53 +00004611 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00004612 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4613 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004614 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4615 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4616 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004617 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00004618</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00004619
Dan Gohmandd8004d2009-07-27 21:53:46 +00004620<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +00004621 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4622 base pointer is not an <i>in bounds</i> address of an allocated object,
4623 or if any of the addresses that would be formed by successive addition of
4624 the offsets implied by the indices to the base address with infinitely
4625 precise arithmetic are not an <i>in bounds</i> address of that allocated
4626 object. The <i>in bounds</i> addresses for an allocated object are all
4627 the addresses that point into the object, plus the address one byte past
4628 the end.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +00004629
4630<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4631 the base address with silently-wrapping two's complement arithmetic, and
4632 the result value of the <tt>getelementptr</tt> may be outside the object
4633 pointed to by the base pointer. The result value may not necessarily be
4634 used to access memory though, even if it happens to point into allocated
4635 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4636 section for more information.</p>
4637
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004638<p>The getelementptr instruction is often confusing. For some more insight into
4639 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +00004640
Chris Lattner7faa8832002-04-14 06:13:44 +00004641<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004642<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004643 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004644 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4645 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +00004646 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004647 <i>; yields i8*:eptr</i>
4648 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +00004649 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +00004650 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004651</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004652
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004653</div>
Reid Spencer47ce1792006-11-09 21:15:49 +00004654
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004655</div>
4656
Chris Lattner00950542001-06-06 20:29:01 +00004657<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004658<h3>
4659 <a name="convertops">Conversion Operations</a>
4660</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004661
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004662<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004663
Reid Spencer2fd21e62006-11-08 01:18:52 +00004664<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004665 which all take a single operand and a type. They perform various bit
4666 conversions on the operand.</p>
4667
Chris Lattner6536cfe2002-05-06 22:08:29 +00004668<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004669<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004670 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004671</h4>
4672
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004673<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004674
4675<h5>Syntax:</h5>
4676<pre>
4677 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4678</pre>
4679
4680<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004681<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4682 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004683
4684<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004685<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
4686 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4687 of the same number of integers.
4688 The bit size of the <tt>value</tt> must be larger than
4689 the bit size of the destination type, <tt>ty2</tt>.
4690 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004691
4692<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004693<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4694 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4695 source size must be larger than the destination size, <tt>trunc</tt> cannot
4696 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004697
4698<h5>Example:</h5>
4699<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004700 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4701 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
4702 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
4703 %W = trunc &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i8&gt; <i>; yields &lt;i8 8, i8 7&gt;</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004704</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004705
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004706</div>
4707
4708<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004709<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004710 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004711</h4>
4712
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004713<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004714
4715<h5>Syntax:</h5>
4716<pre>
4717 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4718</pre>
4719
4720<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004721<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004722 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004723
4724
4725<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +00004726<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
4727 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4728 of the same number of integers.
4729 The bit size of the <tt>value</tt> must be smaller than
4730 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004731 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004732
4733<h5>Semantics:</h5>
4734<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004735 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004736
Reid Spencerb5929522007-01-12 15:46:11 +00004737<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004738
4739<h5>Example:</h5>
4740<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004741 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00004742 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +00004743 %Z = zext &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i32&gt; <i>; yields &lt;i32 8, i32 7&gt;</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004744</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004745
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004746</div>
4747
4748<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004749<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004750 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004751</h4>
4752
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004753<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004754
4755<h5>Syntax:</h5>
4756<pre>
4757 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4758</pre>
4759
4760<h5>Overview:</h5>
4761<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4762
4763<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004764<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
4765 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4766 of the same number of integers.
4767 The bit size of the <tt>value</tt> must be smaller than
4768 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004769 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004770
4771<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004772<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4773 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4774 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004775
Reid Spencerc78f3372007-01-12 03:35:51 +00004776<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004777
4778<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004779<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004780 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00004781 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004782 %Z = sext &lt;2 x i16&gt; &lt;i16 8, i16 7&gt; to &lt;2 x i32&gt; <i>; yields &lt;i32 8, i32 7&gt;</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004783</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004784
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004785</div>
4786
4787<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004788<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004789 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004790</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004791
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004792<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004793
4794<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004795<pre>
4796 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4797</pre>
4798
4799<h5>Overview:</h5>
4800<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004801 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004802
4803<h5>Arguments:</h5>
4804<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004805 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4806 to cast it to. The size of <tt>value</tt> must be larger than the size of
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004807 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004808 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004809
4810<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004811<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004812 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004813 <a href="#t_floating">floating point</a> type. If the value cannot fit
4814 within the destination type, <tt>ty2</tt>, then the results are
4815 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004816
4817<h5>Example:</h5>
4818<pre>
4819 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4820 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4821</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004822
Reid Spencer3fa91b02006-11-09 21:48:10 +00004823</div>
4824
4825<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004826<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004827 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004828</h4>
4829
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004830<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004831
4832<h5>Syntax:</h5>
4833<pre>
4834 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4835</pre>
4836
4837<h5>Overview:</h5>
4838<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004839 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004840
4841<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004842<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004843 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4844 a <a href="#t_floating">floating point</a> type to cast it to. The source
4845 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004846
4847<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00004848<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004849 <a href="#t_floating">floating point</a> type to a larger
4850 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4851 used to make a <i>no-op cast</i> because it always changes bits. Use
4852 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004853
4854<h5>Example:</h5>
4855<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +00004856 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
4857 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004858</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004859
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004860</div>
4861
4862<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004863<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +00004864 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004865</h4>
4866
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004867<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004868
4869<h5>Syntax:</h5>
4870<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +00004871 &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 +00004872</pre>
4873
4874<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +00004875<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004876 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004877
4878<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004879<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4880 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4881 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4882 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4883 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004884
4885<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004886<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004887 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4888 towards zero) unsigned integer value. If the value cannot fit
4889 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004890
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004891<h5>Example:</h5>
4892<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +00004893 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +00004894 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +00004895 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004896</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004897
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004898</div>
4899
4900<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004901<h4>
Reid Spencerd4448792006-11-09 23:03:26 +00004902 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004903</h4>
4904
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004905<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004906
4907<h5>Syntax:</h5>
4908<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00004909 &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 +00004910</pre>
4911
4912<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004913<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004914 <a href="#t_floating">floating point</a> <tt>value</tt> to
4915 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004916
Chris Lattner6536cfe2002-05-06 22:08:29 +00004917<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004918<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4919 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4920 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4921 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4922 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004923
Chris Lattner6536cfe2002-05-06 22:08:29 +00004924<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004925<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004926 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4927 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4928 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004929
Chris Lattner33ba0d92001-07-09 00:26:23 +00004930<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004931<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +00004932 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +00004933 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +00004934 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004935</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004936
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004937</div>
4938
4939<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004940<h4>
Reid Spencerd4448792006-11-09 23:03:26 +00004941 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004942</h4>
4943
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004944<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004945
4946<h5>Syntax:</h5>
4947<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00004948 &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 +00004949</pre>
4950
4951<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00004952<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004953 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004954
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004955<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +00004956<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004957 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4958 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4959 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4960 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004961
4962<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00004963<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004964 integer quantity and converts it to the corresponding floating point
4965 value. If the value cannot fit in the floating point value, the results are
4966 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004967
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004968<h5>Example:</h5>
4969<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004970 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +00004971 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004972</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004973
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004974</div>
4975
4976<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004977<h4>
Reid Spencerd4448792006-11-09 23:03:26 +00004978 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004979</h4>
4980
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004981<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004982
4983<h5>Syntax:</h5>
4984<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00004985 &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 +00004986</pre>
4987
4988<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004989<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4990 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004991
4992<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +00004993<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004994 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4995 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4996 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4997 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004998
4999<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005000<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5001 quantity and converts it to the corresponding floating point value. If the
5002 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005003
5004<h5>Example:</h5>
5005<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00005006 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +00005007 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005008</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005009
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005010</div>
5011
5012<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005013<h4>
Reid Spencer72679252006-11-11 21:00:47 +00005014 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005015</h4>
5016
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005017<div>
Reid Spencer72679252006-11-11 21:00:47 +00005018
5019<h5>Syntax:</h5>
5020<pre>
5021 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5022</pre>
5023
5024<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005025<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5026 the integer type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005027
5028<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005029<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5030 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5031 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005032
5033<h5>Semantics:</h5>
5034<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005035 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5036 truncating or zero extending that value to the size of the integer type. If
5037 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5038 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5039 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5040 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005041
5042<h5>Example:</h5>
5043<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005044 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5045 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +00005046</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005047
Reid Spencer72679252006-11-11 21:00:47 +00005048</div>
5049
5050<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005051<h4>
Reid Spencer72679252006-11-11 21:00:47 +00005052 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005053</h4>
5054
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005055<div>
Reid Spencer72679252006-11-11 21:00:47 +00005056
5057<h5>Syntax:</h5>
5058<pre>
5059 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5060</pre>
5061
5062<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005063<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5064 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005065
5066<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +00005067<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005068 value to cast, and a type to cast it to, which must be a
5069 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005070
5071<h5>Semantics:</h5>
5072<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005073 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5074 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5075 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5076 than the size of a pointer then a zero extension is done. If they are the
5077 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +00005078
5079<h5>Example:</h5>
5080<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005081 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
Gabor Greif6a292012009-10-28 09:21:30 +00005082 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5083 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +00005084</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005085
Reid Spencer72679252006-11-11 21:00:47 +00005086</div>
5087
5088<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005089<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005090 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005091</h4>
5092
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005093<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005094
5095<h5>Syntax:</h5>
5096<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005097 &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 +00005098</pre>
5099
5100<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005101<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005102 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005103
5104<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005105<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5106 non-aggregate first class value, and a type to cast it to, which must also be
5107 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5108 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5109 identical. If the source type is a pointer, the destination type must also be
5110 a pointer. This instruction supports bitwise conversion of vectors to
5111 integers and to vectors of other types (as long as they have the same
5112 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005113
5114<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005115<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005116 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5117 this conversion. The conversion is done as if the <tt>value</tt> had been
5118 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5119 be converted to other pointer types with this instruction. To convert
5120 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5121 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005122
5123<h5>Example:</h5>
5124<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005125 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +00005126 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005127 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00005128</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005129
Misha Brukman9d0919f2003-11-08 01:05:38 +00005130</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005131
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005132</div>
5133
Reid Spencer2fd21e62006-11-08 01:18:52 +00005134<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005135<h3>
5136 <a name="otherops">Other Operations</a>
5137</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005138
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005139<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005140
5141<p>The instructions in this category are the "miscellaneous" instructions, which
5142 defy better classification.</p>
5143
Reid Spencerf3a70a62006-11-18 21:50:54 +00005144<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005145<h4>
5146 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5147</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005148
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005149<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005150
Reid Spencerf3a70a62006-11-18 21:50:54 +00005151<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005152<pre>
5153 &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {i1} or {&lt;N x i1&gt;}:result</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005154</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005155
Reid Spencerf3a70a62006-11-18 21:50:54 +00005156<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005157<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5158 boolean values based on comparison of its two integer, integer vector, or
5159 pointer operands.</p>
5160
Reid Spencerf3a70a62006-11-18 21:50:54 +00005161<h5>Arguments:</h5>
5162<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005163 the condition code indicating the kind of comparison to perform. It is not a
5164 value, just a keyword. The possible condition code are:</p>
5165
Reid Spencerf3a70a62006-11-18 21:50:54 +00005166<ol>
5167 <li><tt>eq</tt>: equal</li>
5168 <li><tt>ne</tt>: not equal </li>
5169 <li><tt>ugt</tt>: unsigned greater than</li>
5170 <li><tt>uge</tt>: unsigned greater or equal</li>
5171 <li><tt>ult</tt>: unsigned less than</li>
5172 <li><tt>ule</tt>: unsigned less or equal</li>
5173 <li><tt>sgt</tt>: signed greater than</li>
5174 <li><tt>sge</tt>: signed greater or equal</li>
5175 <li><tt>slt</tt>: signed less than</li>
5176 <li><tt>sle</tt>: signed less or equal</li>
5177</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005178
Chris Lattner3b19d652007-01-15 01:54:13 +00005179<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005180 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5181 typed. They must also be identical types.</p>
5182
Reid Spencerf3a70a62006-11-18 21:50:54 +00005183<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005184<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5185 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +00005186 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005187 result, as follows:</p>
5188
Reid Spencerf3a70a62006-11-18 21:50:54 +00005189<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005190 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005191 <tt>false</tt> otherwise. No sign interpretation is necessary or
5192 performed.</li>
5193
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005194 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005195 <tt>false</tt> otherwise. No sign interpretation is necessary or
5196 performed.</li>
5197
Reid Spencerf3a70a62006-11-18 21:50:54 +00005198 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005199 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5200
Reid Spencerf3a70a62006-11-18 21:50:54 +00005201 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005202 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5203 to <tt>op2</tt>.</li>
5204
Reid Spencerf3a70a62006-11-18 21:50:54 +00005205 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005206 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5207
Reid Spencerf3a70a62006-11-18 21:50:54 +00005208 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005209 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5210
Reid Spencerf3a70a62006-11-18 21:50:54 +00005211 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005212 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5213
Reid Spencerf3a70a62006-11-18 21:50:54 +00005214 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005215 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5216 to <tt>op2</tt>.</li>
5217
Reid Spencerf3a70a62006-11-18 21:50:54 +00005218 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005219 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5220
Reid Spencerf3a70a62006-11-18 21:50:54 +00005221 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005222 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005223</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005224
Reid Spencerf3a70a62006-11-18 21:50:54 +00005225<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005226 values are compared as if they were integers.</p>
5227
5228<p>If the operands are integer vectors, then they are compared element by
5229 element. The result is an <tt>i1</tt> vector with the same number of elements
5230 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005231
5232<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005233<pre>
5234 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +00005235 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5236 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5237 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5238 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5239 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005240</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +00005241
5242<p>Note that the code generator does not yet support vector types with
5243 the <tt>icmp</tt> instruction.</p>
5244
Reid Spencerf3a70a62006-11-18 21:50:54 +00005245</div>
5246
5247<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005248<h4>
5249 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5250</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005251
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005252<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005253
Reid Spencerf3a70a62006-11-18 21:50:54 +00005254<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005255<pre>
5256 &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {i1} or {&lt;N x i1&gt;}:result</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005257</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005258
Reid Spencerf3a70a62006-11-18 21:50:54 +00005259<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005260<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5261 values based on comparison of its operands.</p>
5262
5263<p>If the operands are floating point scalars, then the result type is a boolean
Nick Lewyckyec38da42009-09-27 00:45:11 +00005264(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005265
5266<p>If the operands are floating point vectors, then the result type is a vector
5267 of boolean with the same number of elements as the operands being
5268 compared.</p>
5269
Reid Spencerf3a70a62006-11-18 21:50:54 +00005270<h5>Arguments:</h5>
5271<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005272 the condition code indicating the kind of comparison to perform. It is not a
5273 value, just a keyword. The possible condition code are:</p>
5274
Reid Spencerf3a70a62006-11-18 21:50:54 +00005275<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +00005276 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005277 <li><tt>oeq</tt>: ordered and equal</li>
5278 <li><tt>ogt</tt>: ordered and greater than </li>
5279 <li><tt>oge</tt>: ordered and greater than or equal</li>
5280 <li><tt>olt</tt>: ordered and less than </li>
5281 <li><tt>ole</tt>: ordered and less than or equal</li>
5282 <li><tt>one</tt>: ordered and not equal</li>
5283 <li><tt>ord</tt>: ordered (no nans)</li>
5284 <li><tt>ueq</tt>: unordered or equal</li>
5285 <li><tt>ugt</tt>: unordered or greater than </li>
5286 <li><tt>uge</tt>: unordered or greater than or equal</li>
5287 <li><tt>ult</tt>: unordered or less than </li>
5288 <li><tt>ule</tt>: unordered or less than or equal</li>
5289 <li><tt>une</tt>: unordered or not equal</li>
5290 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00005291 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005292</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005293
Jeff Cohenb627eab2007-04-29 01:07:00 +00005294<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005295 <i>unordered</i> means that either operand may be a QNAN.</p>
5296
5297<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5298 a <a href="#t_floating">floating point</a> type or
5299 a <a href="#t_vector">vector</a> of floating point type. They must have
5300 identical types.</p>
5301
Reid Spencerf3a70a62006-11-18 21:50:54 +00005302<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +00005303<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005304 according to the condition code given as <tt>cond</tt>. If the operands are
5305 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +00005306 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005307 follows:</p>
5308
Reid Spencerf3a70a62006-11-18 21:50:54 +00005309<ol>
5310 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005311
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005312 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005313 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5314
Reid Spencerb7f26282006-11-19 03:00:14 +00005315 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00005316 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005317
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005318 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005319 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5320
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005321 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005322 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5323
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005324 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005325 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5326
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005327 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005328 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5329
Reid Spencerb7f26282006-11-19 03:00:14 +00005330 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005331
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005332 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005333 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5334
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005335 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005336 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5337
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005338 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005339 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5340
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005341 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005342 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5343
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005344 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005345 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5346
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005347 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005348 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5349
Reid Spencerb7f26282006-11-19 03:00:14 +00005350 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005351
Reid Spencerf3a70a62006-11-18 21:50:54 +00005352 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5353</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005354
5355<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005356<pre>
5357 &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanf72fb672008-09-09 01:02:47 +00005358 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5359 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5360 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005361</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +00005362
5363<p>Note that the code generator does not yet support vector types with
5364 the <tt>fcmp</tt> instruction.</p>
5365
Reid Spencerf3a70a62006-11-18 21:50:54 +00005366</div>
5367
Reid Spencer2fd21e62006-11-08 01:18:52 +00005368<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005369<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00005370 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005371</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00005372
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005373<div>
Chris Lattner5568e942008-05-20 20:48:21 +00005374
Reid Spencer2fd21e62006-11-08 01:18:52 +00005375<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005376<pre>
5377 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5378</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00005379
Reid Spencer2fd21e62006-11-08 01:18:52 +00005380<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005381<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5382 SSA graph representing the function.</p>
5383
Reid Spencer2fd21e62006-11-08 01:18:52 +00005384<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005385<p>The type of the incoming values is specified with the first type field. After
5386 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5387 one pair for each predecessor basic block of the current block. Only values
5388 of <a href="#t_firstclass">first class</a> type may be used as the value
5389 arguments to the PHI node. Only labels may be used as the label
5390 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00005391
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005392<p>There must be no non-phi instructions between the start of a basic block and
5393 the PHI instructions: i.e. PHI instructions must be first in a basic
5394 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00005395
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005396<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5397 occur on the edge from the corresponding predecessor block to the current
5398 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5399 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +00005400
Reid Spencer2fd21e62006-11-08 01:18:52 +00005401<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005402<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005403 specified by the pair corresponding to the predecessor basic block that
5404 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00005405
Reid Spencer2fd21e62006-11-08 01:18:52 +00005406<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00005407<pre>
5408Loop: ; Infinite loop that counts from 0 on up...
5409 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5410 %nextindvar = add i32 %indvar, 1
5411 br label %Loop
5412</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005413
Reid Spencer2fd21e62006-11-08 01:18:52 +00005414</div>
5415
Chris Lattnercc37aae2004-03-12 05:50:16 +00005416<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005417<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005418 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005419</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005420
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005421<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005422
5423<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005424<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +00005425 &lt;result&gt; = select <i>selty</i> &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
5426
Dan Gohman0e451ce2008-10-14 16:51:45 +00005427 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +00005428</pre>
5429
5430<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005431<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5432 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005433
5434
5435<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005436<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5437 values indicating the condition, and two values of the
5438 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5439 vectors and the condition is a scalar, then entire vectors are selected, not
5440 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005441
5442<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005443<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5444 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005445
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005446<p>If the condition is a vector of i1, then the value arguments must be vectors
5447 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005448
5449<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005450<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +00005451 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005452</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +00005453
5454<p>Note that the code generator does not yet support conditions
5455 with vector type.</p>
5456
Chris Lattnercc37aae2004-03-12 05:50:16 +00005457</div>
5458
Robert Bocchino05ccd702006-01-15 20:48:27 +00005459<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005460<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +00005461 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005462</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +00005463
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005464<div>
Chris Lattner2bff5242005-05-06 05:47:36 +00005465
Chris Lattner00950542001-06-06 20:29:01 +00005466<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00005467<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +00005468 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] &lt;ty&gt; [&lt;fnty&gt;*] &lt;fnptrval&gt;(&lt;function args&gt;) [<a href="#fnattrs">fn attrs</a>]
Chris Lattner2bff5242005-05-06 05:47:36 +00005469</pre>
5470
Chris Lattner00950542001-06-06 20:29:01 +00005471<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00005472<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00005473
Chris Lattner00950542001-06-06 20:29:01 +00005474<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00005475<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00005476
Chris Lattner6536cfe2002-05-06 22:08:29 +00005477<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005478 <li>The optional "tail" marker indicates that the callee function does not
5479 access any allocas or varargs in the caller. Note that calls may be
5480 marked "tail" even if they do not occur before
5481 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5482 present, the function call is eligible for tail call optimization,
5483 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +00005484 optimized into a jump</a>. The code generator may optimize calls marked
5485 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5486 sibling call optimization</a> when the caller and callee have
5487 matching signatures, or 2) forced tail call optimization when the
5488 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005489 <ul>
5490 <li>Caller and callee both have the calling
5491 convention <tt>fastcc</tt>.</li>
5492 <li>The call is in tail position (ret immediately follows call and ret
5493 uses value of call or is void).</li>
5494 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +00005495 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005496 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5497 constraints are met.</a></li>
5498 </ul>
5499 </li>
Devang Patelf642f472008-10-06 18:50:38 +00005500
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005501 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5502 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005503 defaults to using C calling conventions. The calling convention of the
5504 call must match the calling convention of the target function, or else the
5505 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +00005506
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005507 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5508 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5509 '<tt>inreg</tt>' attributes are valid here.</li>
5510
5511 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5512 type of the return value. Functions that return no value are marked
5513 <tt><a href="#t_void">void</a></tt>.</li>
5514
5515 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5516 being invoked. The argument types must match the types implied by this
5517 signature. This type can be omitted if the function is not varargs and if
5518 the function type does not return a pointer to a function.</li>
5519
5520 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5521 be invoked. In most cases, this is a direct function invocation, but
5522 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5523 to function value.</li>
5524
5525 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +00005526 signature argument types and parameter attributes. All arguments must be
5527 of <a href="#t_firstclass">first class</a> type. If the function
5528 signature indicates the function accepts a variable number of arguments,
5529 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005530
5531 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5532 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5533 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00005534</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00005535
Chris Lattner00950542001-06-06 20:29:01 +00005536<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005537<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5538 a specified function, with its incoming arguments bound to the specified
5539 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5540 function, control flow continues with the instruction after the function
5541 call, and the return value of the function is bound to the result
5542 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00005543
Chris Lattner00950542001-06-06 20:29:01 +00005544<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00005545<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +00005546 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00005547 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +00005548 %X = tail call i32 @foo() <i>; yields i32</i>
5549 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5550 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +00005551
5552 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +00005553 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +00005554 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5555 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +00005556 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +00005557 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +00005558</pre>
5559
Dale Johannesen07de8d12009-09-24 18:38:21 +00005560<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +00005561standard C99 library as being the C99 library functions, and may perform
5562optimizations or generate code for them under that assumption. This is
5563something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00005564freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +00005565
Misha Brukman9d0919f2003-11-08 01:05:38 +00005566</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005567
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005568<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005569<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +00005570 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005571</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005572
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005573<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005574
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005575<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005576<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005577 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00005578</pre>
5579
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005580<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005581<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005582 the "variable argument" area of a function call. It is used to implement the
5583 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005584
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005585<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005586<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5587 argument. It returns a value of the specified argument type and increments
5588 the <tt>va_list</tt> to point to the next argument. The actual type
5589 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005590
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005591<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005592<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5593 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5594 to the next argument. For more information, see the variable argument
5595 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005596
5597<p>It is legal for this instruction to be called in a function which does not
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005598 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5599 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005600
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005601<p><tt>va_arg</tt> is an LLVM instruction instead of
5602 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5603 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005604
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005605<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005606<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5607
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005608<p>Note that the code generator does not yet fully support va_arg on many
5609 targets. Also, it does not currently support va_arg with aggregate types on
5610 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +00005611
Misha Brukman9d0919f2003-11-08 01:05:38 +00005612</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005613
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005614</div>
5615
5616</div>
5617
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005618<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005619<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +00005620<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00005621
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005622<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005623
5624<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005625 well known names and semantics and are required to follow certain
5626 restrictions. Overall, these intrinsics represent an extension mechanism for
5627 the LLVM language that does not require changing all of the transformations
5628 in LLVM when adding to the language (or the bitcode reader/writer, the
5629 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005630
John Criswellfc6b8952005-05-16 16:17:45 +00005631<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005632 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5633 begin with this prefix. Intrinsic functions must always be external
5634 functions: you cannot define the body of intrinsic functions. Intrinsic
5635 functions may only be used in call or invoke instructions: it is illegal to
5636 take the address of an intrinsic function. Additionally, because intrinsic
5637 functions are part of the LLVM language, it is required if any are added that
5638 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005639
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005640<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5641 family of functions that perform the same operation but on different data
5642 types. Because LLVM can represent over 8 million different integer types,
5643 overloading is used commonly to allow an intrinsic function to operate on any
5644 integer type. One or more of the argument types or the result type can be
5645 overloaded to accept any integer type. Argument types may also be defined as
5646 exactly matching a previous argument's type or the result type. This allows
5647 an intrinsic function which accepts multiple arguments, but needs all of them
5648 to be of the same type, to only be overloaded with respect to a single
5649 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005650
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005651<p>Overloaded intrinsics will have the names of its overloaded argument types
5652 encoded into its function name, each preceded by a period. Only those types
5653 which are overloaded result in a name suffix. Arguments whose type is matched
5654 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5655 can take an integer of any width and returns an integer of exactly the same
5656 integer width. This leads to a family of functions such as
5657 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5658 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5659 suffix is required. Because the argument's type is matched against the return
5660 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +00005661
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005662<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005663 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005664
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005665<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005666<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00005667 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005668</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00005669
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005670<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005671
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005672<p>Variable argument support is defined in LLVM with
5673 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5674 intrinsic functions. These functions are related to the similarly named
5675 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005676
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005677<p>All of these functions operate on arguments that use a target-specific value
5678 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5679 not define what this type is, so all transformations should be prepared to
5680 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005681
Chris Lattner374ab302006-05-15 17:26:46 +00005682<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005683 instruction and the variable argument handling intrinsic functions are
5684 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005685
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00005686<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005687define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +00005688 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +00005689 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +00005690 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005691 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +00005692
5693 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +00005694 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +00005695
5696 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +00005697 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +00005698 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +00005699 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005700 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +00005701
5702 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005703 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +00005704 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +00005705}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005706
5707declare void @llvm.va_start(i8*)
5708declare void @llvm.va_copy(i8*, i8*)
5709declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +00005710</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +00005711
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005712<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005713<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005714 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005715</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +00005716
5717
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005718<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005719
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005720<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005721<pre>
5722 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5723</pre>
5724
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005725<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005726<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5727 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005728
5729<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +00005730<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005731
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005732<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +00005733<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005734 macro available in C. In a target-dependent way, it initializes
5735 the <tt>va_list</tt> element to which the argument points, so that the next
5736 call to <tt>va_arg</tt> will produce the first variable argument passed to
5737 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5738 need to know the last argument of the function as the compiler can figure
5739 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005740
Misha Brukman9d0919f2003-11-08 01:05:38 +00005741</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005742
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005743<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005744<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005745 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005746</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +00005747
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005748<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005749
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005750<h5>Syntax:</h5>
5751<pre>
5752 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5753</pre>
5754
5755<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005756<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005757 which has been initialized previously
5758 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5759 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005760
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005761<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005762<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005763
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005764<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00005765<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005766 macro available in C. In a target-dependent way, it destroys
5767 the <tt>va_list</tt> element to which the argument points. Calls
5768 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5769 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5770 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005771
Misha Brukman9d0919f2003-11-08 01:05:38 +00005772</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005773
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005774<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005775<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005776 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005777</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +00005778
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005779<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005780
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005781<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005782<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00005783 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00005784</pre>
5785
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005786<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005787<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005788 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005789
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005790<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005791<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005792 The second argument is a pointer to a <tt>va_list</tt> element to copy
5793 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005794
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005795<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005796<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005797 macro available in C. In a target-dependent way, it copies the
5798 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5799 element. This intrinsic is necessary because
5800 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5801 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005802
Misha Brukman9d0919f2003-11-08 01:05:38 +00005803</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005804
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005805</div>
5806
Chris Lattner33aec9e2004-02-12 17:01:32 +00005807<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005808<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +00005809 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005810</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +00005811
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005812<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005813
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005814<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +00005815Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005816intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5817roots on the stack</a>, as well as garbage collector implementations that
5818require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5819barriers. Front-ends for type-safe garbage collected languages should generate
5820these intrinsics to make use of the LLVM garbage collectors. For more details,
5821see <a href="GarbageCollection.html">Accurate Garbage Collection with
5822LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +00005823
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005824<p>The garbage collection intrinsics only operate on objects in the generic
5825 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +00005826
Chris Lattnerd7923912004-05-23 21:06:01 +00005827<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005828<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005829 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005830</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +00005831
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005832<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005833
5834<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005835<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00005836 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00005837</pre>
5838
5839<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +00005840<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005841 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005842
5843<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005844<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005845 root pointer. The second pointer (which must be either a constant or a
5846 global value address) contains the meta-data to be associated with the
5847 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005848
5849<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +00005850<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005851 location. At compile-time, the code generator generates information to allow
5852 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5853 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5854 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005855
5856</div>
5857
Chris Lattnerd7923912004-05-23 21:06:01 +00005858<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005859<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005860 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005861</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +00005862
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005863<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005864
5865<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005866<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00005867 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00005868</pre>
5869
5870<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005871<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005872 locations, allowing garbage collector implementations that require read
5873 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005874
5875<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +00005876<p>The second argument is the address to read from, which should be an address
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005877 allocated from the garbage collector. The first object is a pointer to the
5878 start of the referenced object, if needed by the language runtime (otherwise
5879 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005880
5881<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005882<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005883 instruction, but may be replaced with substantially more complex code by the
5884 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5885 may only be used in a function which <a href="#gc">specifies a GC
5886 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005887
5888</div>
5889
Chris Lattnerd7923912004-05-23 21:06:01 +00005890<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005891<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005892 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005893</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +00005894
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005895<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005896
5897<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005898<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00005899 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00005900</pre>
5901
5902<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005903<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005904 locations, allowing garbage collector implementations that require write
5905 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005906
5907<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +00005908<p>The first argument is the reference to store, the second is the start of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005909 object to store it to, and the third is the address of the field of Obj to
5910 store to. If the runtime does not require a pointer to the object, Obj may
5911 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005912
5913<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005914<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005915 instruction, but may be replaced with substantially more complex code by the
5916 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5917 may only be used in a function which <a href="#gc">specifies a GC
5918 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005919
5920</div>
5921
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005922</div>
5923
Chris Lattnerd7923912004-05-23 21:06:01 +00005924<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005925<h3>
Chris Lattner10610642004-02-14 04:08:35 +00005926 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005927</h3>
Chris Lattner10610642004-02-14 04:08:35 +00005928
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005929<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005930
5931<p>These intrinsics are provided by LLVM to expose special features that may
5932 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005933
Chris Lattner10610642004-02-14 04:08:35 +00005934<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005935<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005936 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005937</h4>
Chris Lattner10610642004-02-14 04:08:35 +00005938
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005939<div>
Chris Lattner10610642004-02-14 04:08:35 +00005940
5941<h5>Syntax:</h5>
5942<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00005943 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00005944</pre>
5945
5946<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005947<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5948 target-specific value indicating the return address of the current function
5949 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005950
5951<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005952<p>The argument to this intrinsic indicates which function to return the address
5953 for. Zero indicates the calling function, one indicates its caller, etc.
5954 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005955
5956<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005957<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5958 indicating the return address of the specified call frame, or zero if it
5959 cannot be identified. The value returned by this intrinsic is likely to be
5960 incorrect or 0 for arguments other than zero, so it should only be used for
5961 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005962
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005963<p>Note that calling this intrinsic does not prevent function inlining or other
5964 aggressive transformations, so the value returned may not be that of the
5965 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005966
Chris Lattner10610642004-02-14 04:08:35 +00005967</div>
5968
Chris Lattner10610642004-02-14 04:08:35 +00005969<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005970<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005971 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005972</h4>
Chris Lattner10610642004-02-14 04:08:35 +00005973
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005974<div>
Chris Lattner10610642004-02-14 04:08:35 +00005975
5976<h5>Syntax:</h5>
5977<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00005978 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00005979</pre>
5980
5981<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005982<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5983 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005984
5985<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005986<p>The argument to this intrinsic indicates which function to return the frame
5987 pointer for. Zero indicates the calling function, one indicates its caller,
5988 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005989
5990<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005991<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5992 indicating the frame address of the specified call frame, or zero if it
5993 cannot be identified. The value returned by this intrinsic is likely to be
5994 incorrect or 0 for arguments other than zero, so it should only be used for
5995 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005996
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005997<p>Note that calling this intrinsic does not prevent function inlining or other
5998 aggressive transformations, so the value returned may not be that of the
5999 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006000
Chris Lattner10610642004-02-14 04:08:35 +00006001</div>
6002
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006003<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006004<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006005 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006006</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +00006007
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006008<div>
Chris Lattner57e1f392006-01-13 02:03:13 +00006009
6010<h5>Syntax:</h5>
6011<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006012 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +00006013</pre>
6014
6015<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006016<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6017 of the function stack, for use
6018 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6019 useful for implementing language features like scoped automatic variable
6020 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006021
6022<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006023<p>This intrinsic returns a opaque pointer value that can be passed
6024 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6025 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6026 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6027 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6028 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6029 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006030
6031</div>
6032
6033<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006034<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006035 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006036</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +00006037
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006038<div>
Chris Lattner57e1f392006-01-13 02:03:13 +00006039
6040<h5>Syntax:</h5>
6041<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006042 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +00006043</pre>
6044
6045<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006046<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6047 the function stack to the state it was in when the
6048 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6049 executed. This is useful for implementing language features like scoped
6050 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006051
6052<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006053<p>See the description
6054 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006055
6056</div>
6057
Chris Lattner57e1f392006-01-13 02:03:13 +00006058<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006059<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006060 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006061</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006062
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006063<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006064
6065<h5>Syntax:</h5>
6066<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +00006067 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;, i32 &lt;cache type&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006068</pre>
6069
6070<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006071<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6072 insert a prefetch instruction if supported; otherwise, it is a noop.
6073 Prefetches have no effect on the behavior of the program but can change its
6074 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006075
6076<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006077<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6078 specifier determining if the fetch should be for a read (0) or write (1),
6079 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +00006080 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6081 specifies whether the prefetch is performed on the data (1) or instruction (0)
6082 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6083 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006084
6085<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006086<p>This intrinsic does not modify the behavior of the program. In particular,
6087 prefetches cannot trap and do not produce a value. On targets that support
6088 this intrinsic, the prefetch can provide hints to the processor cache for
6089 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006090
6091</div>
6092
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006093<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006094<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006095 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006096</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006097
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006098<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006099
6100<h5>Syntax:</h5>
6101<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00006102 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006103</pre>
6104
6105<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006106<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6107 Counter (PC) in a region of code to simulators and other tools. The method
6108 is target specific, but it is expected that the marker will use exported
6109 symbols to transmit the PC of the marker. The marker makes no guarantees
6110 that it will remain with any specific instruction after optimizations. It is
6111 possible that the presence of a marker will inhibit optimizations. The
6112 intended use is to be inserted after optimizations to allow correlations of
6113 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006114
6115<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006116<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006117
6118<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006119<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006120 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006121
6122</div>
6123
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006124<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006125<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006126 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006127</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006128
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006129<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006130
6131<h5>Syntax:</h5>
6132<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00006133 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006134</pre>
6135
6136<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006137<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6138 counter register (or similar low latency, high accuracy clocks) on those
6139 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6140 should map to RPCC. As the backing counters overflow quickly (on the order
6141 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006142
6143<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006144<p>When directly supported, reading the cycle counter should not modify any
6145 memory. Implementations are allowed to either return a application specific
6146 value or a system wide value. On backends without support, this is lowered
6147 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006148
6149</div>
6150
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006151</div>
6152
Chris Lattner10610642004-02-14 04:08:35 +00006153<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006154<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006155 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006156</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006157
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006158<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006159
6160<p>LLVM provides intrinsics for a few important standard C library functions.
6161 These intrinsics allow source-language front-ends to pass information about
6162 the alignment of the pointer arguments to the code generator, providing
6163 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006164
Chris Lattner33aec9e2004-02-12 17:01:32 +00006165<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006166<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006167 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006168</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006169
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006170<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006171
6172<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006173<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
Mon P Wange88909b2010-04-07 06:35:53 +00006174 integer bit width and for different address spaces. Not all targets support
6175 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006176
Chris Lattner33aec9e2004-02-12 17:01:32 +00006177<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006178 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006179 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006180 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006181 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00006182</pre>
6183
6184<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006185<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6186 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006187
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006188<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +00006189 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6190 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006191
6192<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006193
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006194<p>The first argument is a pointer to the destination, the second is a pointer
6195 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +00006196 number of bytes to copy, the fourth argument is the alignment of the
6197 source and destination locations, and the fifth is a boolean indicating a
6198 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006199
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006200<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006201 then the caller guarantees that both the source and destination pointers are
6202 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +00006203
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00006204<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6205 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6206 The detailed access behavior is not very cleanly specified and it is unwise
6207 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +00006208
Chris Lattner33aec9e2004-02-12 17:01:32 +00006209<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006210
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006211<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6212 source location to the destination location, which are not allowed to
6213 overlap. It copies "len" bytes of memory over. If the argument is known to
6214 be aligned to some boundary, this can be specified as the fourth argument,
6215 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006216
Chris Lattner33aec9e2004-02-12 17:01:32 +00006217</div>
6218
Chris Lattner0eb51b42004-02-12 18:10:10 +00006219<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006220<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006221 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006222</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006223
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006224<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006225
6226<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +00006227<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Mon P Wange88909b2010-04-07 06:35:53 +00006228 width and for different address space. Not all targets support all bit
6229 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006230
Chris Lattner0eb51b42004-02-12 18:10:10 +00006231<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006232 declare void @llvm.memmove.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006233 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006234 declare void @llvm.memmove.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006235 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00006236</pre>
6237
6238<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006239<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6240 source location to the destination location. It is similar to the
6241 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6242 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006243
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006244<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +00006245 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6246 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006247
6248<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006249
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006250<p>The first argument is a pointer to the destination, the second is a pointer
6251 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +00006252 number of bytes to copy, the fourth argument is the alignment of the
6253 source and destination locations, and the fifth is a boolean indicating a
6254 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006255
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006256<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006257 then the caller guarantees that the source and destination pointers are
6258 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +00006259
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00006260<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6261 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6262 The detailed access behavior is not very cleanly specified and it is unwise
6263 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +00006264
Chris Lattner0eb51b42004-02-12 18:10:10 +00006265<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006266
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006267<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6268 source location to the destination location, which may overlap. It copies
6269 "len" bytes of memory over. If the argument is known to be aligned to some
6270 boundary, this can be specified as the fourth argument, otherwise it should
6271 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006272
Chris Lattner0eb51b42004-02-12 18:10:10 +00006273</div>
6274
Chris Lattner10610642004-02-14 04:08:35 +00006275<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006276<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006277 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006278</h4>
Chris Lattner10610642004-02-14 04:08:35 +00006279
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006280<div>
Chris Lattner10610642004-02-14 04:08:35 +00006281
6282<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +00006283<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
John Criswellcdcbbfc2010-07-30 16:30:28 +00006284 width and for different address spaces. However, not all targets support all
6285 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006286
Chris Lattner10610642004-02-14 04:08:35 +00006287<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006288 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +00006289 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006290 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +00006291 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00006292</pre>
6293
6294<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006295<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6296 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006297
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006298<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +00006299 intrinsic does not return a value and takes extra alignment/volatile
6300 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006301
6302<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006303<p>The first argument is a pointer to the destination to fill, the second is the
John Criswellcdcbbfc2010-07-30 16:30:28 +00006304 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006305 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +00006306 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006307
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006308<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006309 then the caller guarantees that the destination pointer is aligned to that
6310 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006311
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00006312<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6313 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6314 The detailed access behavior is not very cleanly specified and it is unwise
6315 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +00006316
Chris Lattner10610642004-02-14 04:08:35 +00006317<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006318<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6319 at the destination location. If the argument is known to be aligned to some
6320 boundary, this can be specified as the fourth argument, otherwise it should
6321 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006322
Chris Lattner10610642004-02-14 04:08:35 +00006323</div>
6324
Chris Lattner32006282004-06-11 02:28:03 +00006325<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006326<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006327 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006328</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +00006329
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006330<div>
Chris Lattnera4d74142005-07-21 01:29:16 +00006331
6332<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006333<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6334 floating point or vector of floating point type. Not all targets support all
6335 types however.</p>
6336
Chris Lattnera4d74142005-07-21 01:29:16 +00006337<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +00006338 declare float @llvm.sqrt.f32(float %Val)
6339 declare double @llvm.sqrt.f64(double %Val)
6340 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6341 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6342 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +00006343</pre>
6344
6345<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006346<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6347 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6348 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6349 behavior for negative numbers other than -0.0 (which allows for better
6350 optimization, because there is no need to worry about errno being
6351 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +00006352
6353<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006354<p>The argument and return value are floating point numbers of the same
6355 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +00006356
6357<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006358<p>This function returns the sqrt of the specified operand if it is a
6359 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +00006360
Chris Lattnera4d74142005-07-21 01:29:16 +00006361</div>
6362
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006363<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006364<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006365 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006366</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006367
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006368<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006369
6370<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006371<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6372 floating point or vector of floating point type. Not all targets support all
6373 types however.</p>
6374
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006375<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +00006376 declare float @llvm.powi.f32(float %Val, i32 %power)
6377 declare double @llvm.powi.f64(double %Val, i32 %power)
6378 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6379 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6380 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006381</pre>
6382
6383<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006384<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6385 specified (positive or negative) power. The order of evaluation of
6386 multiplications is not defined. When a vector of floating point type is
6387 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006388
6389<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006390<p>The second argument is an integer power, and the first is a value to raise to
6391 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006392
6393<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006394<p>This function returns the first value raised to the second power with an
6395 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006396
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006397</div>
6398
Dan Gohman91c284c2007-10-15 20:30:11 +00006399<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006400<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006401 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006402</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006403
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006404<div>
Dan Gohman91c284c2007-10-15 20:30:11 +00006405
6406<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006407<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6408 floating point or vector of floating point type. Not all targets support all
6409 types however.</p>
6410
Dan Gohman91c284c2007-10-15 20:30:11 +00006411<pre>
6412 declare float @llvm.sin.f32(float %Val)
6413 declare double @llvm.sin.f64(double %Val)
6414 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6415 declare fp128 @llvm.sin.f128(fp128 %Val)
6416 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6417</pre>
6418
6419<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006420<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006421
6422<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006423<p>The argument and return value are floating point numbers of the same
6424 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006425
6426<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006427<p>This function returns the sine of the specified operand, returning the same
6428 values as the libm <tt>sin</tt> functions would, and handles error conditions
6429 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006430
Dan Gohman91c284c2007-10-15 20:30:11 +00006431</div>
6432
6433<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006434<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006435 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006436</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006437
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006438<div>
Dan Gohman91c284c2007-10-15 20:30:11 +00006439
6440<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006441<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6442 floating point or vector of floating point type. Not all targets support all
6443 types however.</p>
6444
Dan Gohman91c284c2007-10-15 20:30:11 +00006445<pre>
6446 declare float @llvm.cos.f32(float %Val)
6447 declare double @llvm.cos.f64(double %Val)
6448 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6449 declare fp128 @llvm.cos.f128(fp128 %Val)
6450 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6451</pre>
6452
6453<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006454<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006455
6456<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006457<p>The argument and return value are floating point numbers of the same
6458 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006459
6460<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006461<p>This function returns the cosine of the specified operand, returning the same
6462 values as the libm <tt>cos</tt> functions would, and handles error conditions
6463 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006464
Dan Gohman91c284c2007-10-15 20:30:11 +00006465</div>
6466
6467<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006468<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006469 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006470</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006471
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006472<div>
Dan Gohman91c284c2007-10-15 20:30:11 +00006473
6474<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006475<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6476 floating point or vector of floating point type. Not all targets support all
6477 types however.</p>
6478
Dan Gohman91c284c2007-10-15 20:30:11 +00006479<pre>
6480 declare float @llvm.pow.f32(float %Val, float %Power)
6481 declare double @llvm.pow.f64(double %Val, double %Power)
6482 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6483 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6484 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6485</pre>
6486
6487<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006488<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6489 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006490
6491<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006492<p>The second argument is a floating point power, and the first is a value to
6493 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006494
6495<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006496<p>This function returns the first value raised to the second power, returning
6497 the same values as the libm <tt>pow</tt> functions would, and handles error
6498 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006499
Dan Gohman91c284c2007-10-15 20:30:11 +00006500</div>
6501
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006502</div>
6503
Dan Gohman4e9011c2011-05-23 21:13:03 +00006504<!-- _______________________________________________________________________ -->
6505<h4>
6506 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
6507</h4>
6508
6509<div>
6510
6511<h5>Syntax:</h5>
6512<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
6513 floating point or vector of floating point type. Not all targets support all
6514 types however.</p>
6515
6516<pre>
6517 declare float @llvm.exp.f32(float %Val)
6518 declare double @llvm.exp.f64(double %Val)
6519 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
6520 declare fp128 @llvm.exp.f128(fp128 %Val)
6521 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
6522</pre>
6523
6524<h5>Overview:</h5>
6525<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
6526
6527<h5>Arguments:</h5>
6528<p>The argument and return value are floating point numbers of the same
6529 type.</p>
6530
6531<h5>Semantics:</h5>
6532<p>This function returns the same values as the libm <tt>exp</tt> functions
6533 would, and handles error conditions in the same way.</p>
6534
6535</div>
6536
6537<!-- _______________________________________________________________________ -->
6538<h4>
6539 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
6540</h4>
6541
6542<div>
6543
6544<h5>Syntax:</h5>
6545<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
6546 floating point or vector of floating point type. Not all targets support all
6547 types however.</p>
6548
6549<pre>
6550 declare float @llvm.log.f32(float %Val)
6551 declare double @llvm.log.f64(double %Val)
6552 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
6553 declare fp128 @llvm.log.f128(fp128 %Val)
6554 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
6555</pre>
6556
6557<h5>Overview:</h5>
6558<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
6559
6560<h5>Arguments:</h5>
6561<p>The argument and return value are floating point numbers of the same
6562 type.</p>
6563
6564<h5>Semantics:</h5>
6565<p>This function returns the same values as the libm <tt>log</tt> functions
6566 would, and handles error conditions in the same way.</p>
6567
6568</div>
6569
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006570<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006571<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +00006572 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006573</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006574
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006575<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006576
6577<p>LLVM provides intrinsics for a few important bit manipulation operations.
6578 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006579
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006580<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006581<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006582 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006583</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +00006584
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006585<div>
Nate Begeman7e36c472006-01-13 23:26:38 +00006586
6587<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00006588<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006589 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6590
Nate Begeman7e36c472006-01-13 23:26:38 +00006591<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00006592 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6593 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6594 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +00006595</pre>
6596
6597<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006598<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6599 values with an even number of bytes (positive multiple of 16 bits). These
6600 are useful for performing operations on data that is not in the target's
6601 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +00006602
6603<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006604<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6605 and low byte of the input i16 swapped. Similarly,
6606 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6607 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6608 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6609 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6610 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6611 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +00006612
6613</div>
6614
6615<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006616<h4>
Reid Spencer0b118202006-01-16 21:12:35 +00006617 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006618</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006619
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006620<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006621
6622<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00006623<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006624 width. Not all targets support all bit widths however.</p>
6625
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006626<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006627 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006628 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00006629 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006630 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6631 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006632</pre>
6633
6634<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006635<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6636 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006637
6638<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006639<p>The only argument is the value to be counted. The argument may be of any
6640 integer type. The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006641
6642<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006643<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006644
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006645</div>
6646
6647<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006648<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +00006649 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006650</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006651
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006652<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006653
6654<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006655<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6656 integer bit width. Not all targets support all bit widths however.</p>
6657
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006658<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00006659 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6660 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00006661 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006662 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6663 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006664</pre>
6665
6666<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006667<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6668 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006669
6670<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006671<p>The only argument is the value to be counted. The argument may be of any
6672 integer type. The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006673
6674<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006675<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6676 zeros in a variable. If the src == 0 then the result is the size in bits of
6677 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006678
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006679</div>
Chris Lattner32006282004-06-11 02:28:03 +00006680
Chris Lattnereff29ab2005-05-15 19:39:26 +00006681<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006682<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +00006683 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006684</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006685
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006686<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006687
6688<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006689<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6690 integer bit width. Not all targets support all bit widths however.</p>
6691
Chris Lattnereff29ab2005-05-15 19:39:26 +00006692<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00006693 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6694 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00006695 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006696 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6697 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +00006698</pre>
6699
6700<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006701<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6702 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006703
6704<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006705<p>The only argument is the value to be counted. The argument may be of any
6706 integer type. The return type must match the argument type.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006707
6708<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006709<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6710 zeros in a variable. If the src == 0 then the result is the size in bits of
6711 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006712
Chris Lattnereff29ab2005-05-15 19:39:26 +00006713</div>
6714
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006715</div>
6716
Bill Wendlingda01af72009-02-08 04:04:40 +00006717<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006718<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +00006719 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006720</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +00006721
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006722<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006723
6724<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +00006725
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006726<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006727<h4>
6728 <a name="int_sadd_overflow">
6729 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
6730 </a>
6731</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006732
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006733<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006734
6735<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006736<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006737 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006738
6739<pre>
6740 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6741 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6742 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6743</pre>
6744
6745<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006746<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006747 a signed addition of the two arguments, and indicate whether an overflow
6748 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006749
6750<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006751<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006752 be of integer types of any bit width, but they must have the same bit
6753 width. The second element of the result structure must be of
6754 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6755 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006756
6757<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006758<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006759 a signed addition of the two variables. They return a structure &mdash; the
6760 first element of which is the signed summation, and the second element of
6761 which is a bit specifying if the signed summation resulted in an
6762 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006763
6764<h5>Examples:</h5>
6765<pre>
6766 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6767 %sum = extractvalue {i32, i1} %res, 0
6768 %obit = extractvalue {i32, i1} %res, 1
6769 br i1 %obit, label %overflow, label %normal
6770</pre>
6771
6772</div>
6773
6774<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006775<h4>
6776 <a name="int_uadd_overflow">
6777 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
6778 </a>
6779</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006780
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006781<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006782
6783<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006784<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006785 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006786
6787<pre>
6788 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6789 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6790 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6791</pre>
6792
6793<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006794<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006795 an unsigned addition of the two arguments, and indicate whether a carry
6796 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006797
6798<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006799<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006800 be of integer types of any bit width, but they must have the same bit
6801 width. The second element of the result structure must be of
6802 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6803 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006804
6805<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006806<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006807 an unsigned addition of the two arguments. They return a structure &mdash;
6808 the first element of which is the sum, and the second element of which is a
6809 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006810
6811<h5>Examples:</h5>
6812<pre>
6813 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6814 %sum = extractvalue {i32, i1} %res, 0
6815 %obit = extractvalue {i32, i1} %res, 1
6816 br i1 %obit, label %carry, label %normal
6817</pre>
6818
6819</div>
6820
6821<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006822<h4>
6823 <a name="int_ssub_overflow">
6824 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
6825 </a>
6826</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006827
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006828<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006829
6830<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006831<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006832 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006833
6834<pre>
6835 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6836 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6837 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6838</pre>
6839
6840<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006841<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006842 a signed subtraction of the two arguments, and indicate whether an overflow
6843 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006844
6845<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006846<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006847 be of integer types of any bit width, but they must have the same bit
6848 width. The second element of the result structure must be of
6849 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6850 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006851
6852<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006853<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006854 a signed subtraction of the two arguments. They return a structure &mdash;
6855 the first element of which is the subtraction, and the second element of
6856 which is a bit specifying if the signed subtraction resulted in an
6857 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006858
6859<h5>Examples:</h5>
6860<pre>
6861 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6862 %sum = extractvalue {i32, i1} %res, 0
6863 %obit = extractvalue {i32, i1} %res, 1
6864 br i1 %obit, label %overflow, label %normal
6865</pre>
6866
6867</div>
6868
6869<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006870<h4>
6871 <a name="int_usub_overflow">
6872 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
6873 </a>
6874</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006875
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006876<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006877
6878<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006879<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006880 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006881
6882<pre>
6883 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6884 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6885 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6886</pre>
6887
6888<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006889<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006890 an unsigned subtraction of the two arguments, and indicate whether an
6891 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006892
6893<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006894<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006895 be of integer types of any bit width, but they must have the same bit
6896 width. The second element of the result structure must be of
6897 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6898 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006899
6900<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006901<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006902 an unsigned subtraction of the two arguments. They return a structure &mdash;
6903 the first element of which is the subtraction, and the second element of
6904 which is a bit specifying if the unsigned subtraction resulted in an
6905 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006906
6907<h5>Examples:</h5>
6908<pre>
6909 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6910 %sum = extractvalue {i32, i1} %res, 0
6911 %obit = extractvalue {i32, i1} %res, 1
6912 br i1 %obit, label %overflow, label %normal
6913</pre>
6914
6915</div>
6916
6917<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006918<h4>
6919 <a name="int_smul_overflow">
6920 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
6921 </a>
6922</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006923
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006924<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006925
6926<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006927<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006928 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006929
6930<pre>
6931 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6932 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6933 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6934</pre>
6935
6936<h5>Overview:</h5>
6937
6938<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006939 a signed multiplication of the two arguments, and indicate whether an
6940 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006941
6942<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006943<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006944 be of integer types of any bit width, but they must have the same bit
6945 width. The second element of the result structure must be of
6946 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6947 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006948
6949<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006950<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006951 a signed multiplication of the two arguments. They return a structure &mdash;
6952 the first element of which is the multiplication, and the second element of
6953 which is a bit specifying if the signed multiplication resulted in an
6954 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006955
6956<h5>Examples:</h5>
6957<pre>
6958 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6959 %sum = extractvalue {i32, i1} %res, 0
6960 %obit = extractvalue {i32, i1} %res, 1
6961 br i1 %obit, label %overflow, label %normal
6962</pre>
6963
Reid Spencerf86037f2007-04-11 23:23:49 +00006964</div>
6965
Bill Wendling41b485c2009-02-08 23:00:09 +00006966<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006967<h4>
6968 <a name="int_umul_overflow">
6969 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
6970 </a>
6971</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +00006972
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006973<div>
Bill Wendling41b485c2009-02-08 23:00:09 +00006974
6975<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00006976<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006977 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00006978
6979<pre>
6980 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6981 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6982 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6983</pre>
6984
6985<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00006986<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006987 a unsigned multiplication of the two arguments, and indicate whether an
6988 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00006989
6990<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00006991<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006992 be of integer types of any bit width, but they must have the same bit
6993 width. The second element of the result structure must be of
6994 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6995 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00006996
6997<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00006998<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006999 an unsigned multiplication of the two arguments. They return a structure
7000 &mdash; the first element of which is the multiplication, and the second
7001 element of which is a bit specifying if the unsigned multiplication resulted
7002 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00007003
7004<h5>Examples:</h5>
7005<pre>
7006 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7007 %sum = extractvalue {i32, i1} %res, 0
7008 %obit = extractvalue {i32, i1} %res, 1
7009 br i1 %obit, label %overflow, label %normal
7010</pre>
7011
7012</div>
7013
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007014</div>
7015
Chris Lattner8ff75902004-01-06 05:31:32 +00007016<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007017<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007018 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007019</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007020
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007021<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007022
Chris Lattner0cec9c82010-03-15 04:12:21 +00007023<p>Half precision floating point is a storage-only format. This means that it is
7024 a dense encoding (in memory) but does not support computation in the
7025 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +00007026
Chris Lattner0cec9c82010-03-15 04:12:21 +00007027<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +00007028 value as an i16, then convert it to float with <a
7029 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7030 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +00007031 double etc). To store the value back to memory, it is first converted to
7032 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +00007033 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7034 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007035
7036<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007037<h4>
7038 <a name="int_convert_to_fp16">
7039 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7040 </a>
7041</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007042
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007043<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007044
7045<h5>Syntax:</h5>
7046<pre>
7047 declare i16 @llvm.convert.to.fp16(f32 %a)
7048</pre>
7049
7050<h5>Overview:</h5>
7051<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7052 a conversion from single precision floating point format to half precision
7053 floating point format.</p>
7054
7055<h5>Arguments:</h5>
7056<p>The intrinsic function contains single argument - the value to be
7057 converted.</p>
7058
7059<h5>Semantics:</h5>
7060<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7061 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +00007062 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +00007063 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007064
7065<h5>Examples:</h5>
7066<pre>
7067 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7068 store i16 %res, i16* @x, align 2
7069</pre>
7070
7071</div>
7072
7073<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007074<h4>
7075 <a name="int_convert_from_fp16">
7076 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7077 </a>
7078</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007079
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007080<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007081
7082<h5>Syntax:</h5>
7083<pre>
7084 declare f32 @llvm.convert.from.fp16(i16 %a)
7085</pre>
7086
7087<h5>Overview:</h5>
7088<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7089 a conversion from half precision floating point format to single precision
7090 floating point format.</p>
7091
7092<h5>Arguments:</h5>
7093<p>The intrinsic function contains single argument - the value to be
7094 converted.</p>
7095
7096<h5>Semantics:</h5>
7097<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +00007098 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +00007099 precision floating point format. The input half-float value is represented by
7100 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007101
7102<h5>Examples:</h5>
7103<pre>
7104 %a = load i16* @x, align 2
7105 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7106</pre>
7107
7108</div>
7109
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007110</div>
7111
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007112<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007113<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00007114 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007115</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00007116
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007117<div>
Chris Lattner8ff75902004-01-06 05:31:32 +00007118
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007119<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7120 prefix), are described in
7121 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7122 Level Debugging</a> document.</p>
7123
7124</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00007125
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007126<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007127<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007128 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007129</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007130
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007131<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007132
7133<p>The LLVM exception handling intrinsics (which all start with
7134 <tt>llvm.eh.</tt> prefix), are described in
7135 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7136 Handling</a> document.</p>
7137
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007138</div>
7139
Tanya Lattner6d806e92007-06-15 20:50:54 +00007140<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007141<h3>
Duncan Sandsf7331b32007-09-11 14:10:23 +00007142 <a name="int_trampoline">Trampoline Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007143</h3>
Duncan Sands36397f52007-07-27 12:58:54 +00007144
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007145<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007146
7147<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +00007148 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7149 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007150 function pointer lacking the nest parameter - the caller does not need to
7151 provide a value for it. Instead, the value to use is stored in advance in a
7152 "trampoline", a block of memory usually allocated on the stack, which also
7153 contains code to splice the nest value into the argument list. This is used
7154 to implement the GCC nested function address extension.</p>
7155
7156<p>For example, if the function is
7157 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7158 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7159 follows:</p>
7160
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00007161<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +00007162 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7163 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007164 %p = call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval)
Duncan Sandsf7331b32007-09-11 14:10:23 +00007165 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +00007166</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007167
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007168<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7169 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007170
Duncan Sands36397f52007-07-27 12:58:54 +00007171<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007172<h4>
7173 <a name="int_it">
7174 '<tt>llvm.init.trampoline</tt>' Intrinsic
7175 </a>
7176</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007177
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007178<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007179
Duncan Sands36397f52007-07-27 12:58:54 +00007180<h5>Syntax:</h5>
7181<pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007182 declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands36397f52007-07-27 12:58:54 +00007183</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007184
Duncan Sands36397f52007-07-27 12:58:54 +00007185<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007186<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
7187 function pointer suitable for executing it.</p>
7188
Duncan Sands36397f52007-07-27 12:58:54 +00007189<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007190<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7191 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7192 sufficiently aligned block of memory; this memory is written to by the
7193 intrinsic. Note that the size and the alignment are target-specific - LLVM
7194 currently provides no portable way of determining them, so a front-end that
7195 generates this intrinsic needs to have some target-specific knowledge.
7196 The <tt>func</tt> argument must hold a function bitcast to
7197 an <tt>i8*</tt>.</p>
7198
Duncan Sands36397f52007-07-27 12:58:54 +00007199<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007200<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7201 dependent code, turning it into a function. A pointer to this function is
7202 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7203 function pointer type</a> before being called. The new function's signature
7204 is the same as that of <tt>func</tt> with any arguments marked with
7205 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7206 is allowed, and it must be of pointer type. Calling the new function is
7207 equivalent to calling <tt>func</tt> with the same argument list, but
7208 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7209 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7210 by <tt>tramp</tt> is modified, then the effect of any later call to the
7211 returned function pointer is undefined.</p>
7212
Duncan Sands36397f52007-07-27 12:58:54 +00007213</div>
7214
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007215</div>
7216
Duncan Sands36397f52007-07-27 12:58:54 +00007217<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007218<h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007219 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007220</h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007221
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007222<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007223
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007224<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7225 hardware constructs for atomic operations and memory synchronization. This
7226 provides an interface to the hardware, not an interface to the programmer. It
7227 is aimed at a low enough level to allow any programming models or APIs
7228 (Application Programming Interfaces) which need atomic behaviors to map
7229 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7230 hardware provides a "universal IR" for source languages, it also provides a
7231 starting point for developing a "universal" atomic operation and
7232 synchronization IR.</p>
7233
7234<p>These do <em>not</em> form an API such as high-level threading libraries,
7235 software transaction memory systems, atomic primitives, and intrinsic
7236 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7237 application libraries. The hardware interface provided by LLVM should allow
7238 a clean implementation of all of these APIs and parallel programming models.
7239 No one model or paradigm should be selected above others unless the hardware
7240 itself ubiquitously does so.</p>
7241
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007242<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007243<h4>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007244 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007245</h4>
7246
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007247<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007248<h5>Syntax:</h5>
7249<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007250 declare void @llvm.memory.barrier(i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;, i1 &lt;device&gt;)
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007251</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007252
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007253<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007254<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7255 specific pairs of memory access types.</p>
7256
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007257<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007258<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7259 The first four arguments enables a specific barrier as listed below. The
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00007260 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007261 memory.</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007262
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007263<ul>
7264 <li><tt>ll</tt>: load-load barrier</li>
7265 <li><tt>ls</tt>: load-store barrier</li>
7266 <li><tt>sl</tt>: store-load barrier</li>
7267 <li><tt>ss</tt>: store-store barrier</li>
7268 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7269</ul>
7270
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007271<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007272<p>This intrinsic causes the system to enforce some ordering constraints upon
7273 the loads and stores of the program. This barrier does not
7274 indicate <em>when</em> any events will occur, it only enforces
7275 an <em>order</em> in which they occur. For any of the specified pairs of load
7276 and store operations (f.ex. load-load, or store-load), all of the first
7277 operations preceding the barrier will complete before any of the second
7278 operations succeeding the barrier begin. Specifically the semantics for each
7279 pairing is as follows:</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007280
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007281<ul>
7282 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7283 after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007284 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007285 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007286 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007287 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007288 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007289 load after the barrier begins.</li>
7290</ul>
7291
7292<p>These semantics are applied with a logical "and" behavior when more than one
7293 is enabled in a single memory barrier intrinsic.</p>
7294
7295<p>Backends may implement stronger barriers than those requested when they do
7296 not support as fine grained a barrier as requested. Some architectures do
7297 not need all types of barriers and on such architectures, these become
7298 noops.</p>
7299
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007300<h5>Example:</h5>
7301<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007302%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7303%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007304 store i32 4, %ptr
7305
7306%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007307 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false)
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007308 <i>; guarantee the above finishes</i>
7309 store i32 8, %ptr <i>; before this begins</i>
7310</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007311
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007312</div>
7313
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007314<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007315<h4>
Mon P Wang28873102008-06-25 08:15:39 +00007316 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007317</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007318
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007319<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007320
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007321<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007322<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7323 any integer bit width and for different address spaces. Not all targets
7324 support all bit widths however.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007325
7326<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007327 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7328 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7329 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7330 declare i64 @llvm.atomic.cmp.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007331</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007332
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007333<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007334<p>This loads a value in memory and compares it to a given value. If they are
7335 equal, it stores a new value into the memory.</p>
7336
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007337<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007338<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7339 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7340 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7341 this integer type. While any bit width integer may be used, targets may only
7342 lower representations they support in hardware.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007343
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007344<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007345<p>This entire intrinsic must be executed atomically. It first loads the value
7346 in memory pointed to by <tt>ptr</tt> and compares it with the
7347 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7348 memory. The loaded value is yielded in all cases. This provides the
7349 equivalent of an atomic compare-and-swap operation within the SSA
7350 framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007351
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007352<h5>Examples:</h5>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007353<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007354%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7355%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007356 store i32 4, %ptr
7357
7358%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007359%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007360 <i>; yields {i32}:result1 = 4</i>
7361%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7362%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7363
7364%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007365%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007366 <i>; yields {i32}:result2 = 8</i>
7367%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7368
7369%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7370</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007371
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007372</div>
7373
7374<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007375<h4>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007376 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007377</h4>
7378
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007379<div>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007380<h5>Syntax:</h5>
7381
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007382<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7383 integer bit width. Not all targets support all bit widths however.</p>
7384
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007385<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007386 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7387 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7388 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7389 declare i64 @llvm.atomic.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;val&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007390</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007391
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007392<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007393<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7394 the value from memory. It then stores the value in <tt>val</tt> in the memory
7395 at <tt>ptr</tt>.</p>
7396
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007397<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007398<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7399 the <tt>val</tt> argument and the result must be integers of the same bit
7400 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7401 integer type. The targets may only lower integer representations they
7402 support.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007403
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007404<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007405<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7406 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7407 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007408
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007409<h5>Examples:</h5>
7410<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007411%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7412%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007413 store i32 4, %ptr
7414
7415%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007416%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007417 <i>; yields {i32}:result1 = 4</i>
7418%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7419%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7420
7421%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007422%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007423 <i>; yields {i32}:result2 = 8</i>
7424
7425%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7426%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7427</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007428
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007429</div>
7430
7431<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007432<h4>
Mon P Wang28873102008-06-25 08:15:39 +00007433 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007434</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007435
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007436<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007437
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007438<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007439<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7440 any integer bit width. Not all targets support all bit widths however.</p>
7441
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007442<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007443 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7444 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7445 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7446 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007447</pre>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007448
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007449<h5>Overview:</h5>
7450<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7451 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7452
7453<h5>Arguments:</h5>
7454<p>The intrinsic takes two arguments, the first a pointer to an integer value
7455 and the second an integer value. The result is also an integer value. These
7456 integer types can have any bit width, but they must all have the same bit
7457 width. The targets may only lower integer representations they support.</p>
7458
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007459<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007460<p>This intrinsic does a series of operations atomically. It first loads the
7461 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7462 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007463
7464<h5>Examples:</h5>
7465<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007466%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7467%ptr = bitcast i8* %mallocP to i32*
7468 store i32 4, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007469%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007470 <i>; yields {i32}:result1 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007471%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007472 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007473%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007474 <i>; yields {i32}:result3 = 10</i>
Mon P Wang28873102008-06-25 08:15:39 +00007475%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007476</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007477
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007478</div>
7479
Mon P Wang28873102008-06-25 08:15:39 +00007480<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007481<h4>
Mon P Wang28873102008-06-25 08:15:39 +00007482 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007483</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007484
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007485<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007486
Mon P Wang28873102008-06-25 08:15:39 +00007487<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007488<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7489 any integer bit width and for different address spaces. Not all targets
7490 support all bit widths however.</p>
7491
Mon P Wang28873102008-06-25 08:15:39 +00007492<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007493 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7494 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7495 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7496 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007497</pre>
Mon P Wang28873102008-06-25 08:15:39 +00007498
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007499<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007500<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007501 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7502
7503<h5>Arguments:</h5>
7504<p>The intrinsic takes two arguments, the first a pointer to an integer value
7505 and the second an integer value. The result is also an integer value. These
7506 integer types can have any bit width, but they must all have the same bit
7507 width. The targets may only lower integer representations they support.</p>
7508
Mon P Wang28873102008-06-25 08:15:39 +00007509<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007510<p>This intrinsic does a series of operations atomically. It first loads the
7511 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7512 result to <tt>ptr</tt>. It yields the original value stored
7513 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007514
7515<h5>Examples:</h5>
7516<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007517%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7518%ptr = bitcast i8* %mallocP to i32*
7519 store i32 8, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007520%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang28873102008-06-25 08:15:39 +00007521 <i>; yields {i32}:result1 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007522%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang28873102008-06-25 08:15:39 +00007523 <i>; yields {i32}:result2 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007524%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang28873102008-06-25 08:15:39 +00007525 <i>; yields {i32}:result3 = 2</i>
7526%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7527</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007528
Mon P Wang28873102008-06-25 08:15:39 +00007529</div>
7530
7531<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007532<h4>
7533 <a name="int_atomic_load_and">
7534 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
7535 </a>
7536 <br>
7537 <a name="int_atomic_load_nand">
7538 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
7539 </a>
7540 <br>
7541 <a name="int_atomic_load_or">
7542 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
7543 </a>
7544 <br>
7545 <a name="int_atomic_load_xor">
7546 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
7547 </a>
7548</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007549
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007550<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007551
Mon P Wang28873102008-06-25 08:15:39 +00007552<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007553<p>These are overloaded intrinsics. You can
7554 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7555 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7556 bit width and for different address spaces. Not all targets support all bit
7557 widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007558
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007559<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007560 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7561 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7562 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7563 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007564</pre>
7565
7566<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007567 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7568 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7569 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7570 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007571</pre>
7572
7573<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007574 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7575 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7576 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7577 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007578</pre>
7579
7580<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007581 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7582 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7583 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7584 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007585</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007586
Mon P Wang28873102008-06-25 08:15:39 +00007587<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007588<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7589 the value stored in memory at <tt>ptr</tt>. It yields the original value
7590 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007591
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007592<h5>Arguments:</h5>
7593<p>These intrinsics take two arguments, the first a pointer to an integer value
7594 and the second an integer value. The result is also an integer value. These
7595 integer types can have any bit width, but they must all have the same bit
7596 width. The targets may only lower integer representations they support.</p>
7597
Mon P Wang28873102008-06-25 08:15:39 +00007598<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007599<p>These intrinsics does a series of operations atomically. They first load the
7600 value stored at <tt>ptr</tt>. They then do the bitwise
7601 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7602 original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007603
7604<h5>Examples:</h5>
7605<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007606%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7607%ptr = bitcast i8* %mallocP to i32*
7608 store i32 0x0F0F, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007609%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +00007610 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007611%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +00007612 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007613%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +00007614 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007615%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +00007616 <i>; yields {i32}:result3 = FF</i>
7617%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7618</pre>
Mon P Wang28873102008-06-25 08:15:39 +00007619
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007620</div>
Mon P Wang28873102008-06-25 08:15:39 +00007621
7622<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007623<h4>
7624 <a name="int_atomic_load_max">
7625 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
7626 </a>
7627 <br>
7628 <a name="int_atomic_load_min">
7629 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
7630 </a>
7631 <br>
7632 <a name="int_atomic_load_umax">
7633 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
7634 </a>
7635 <br>
7636 <a name="int_atomic_load_umin">
7637 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
7638 </a>
7639</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007640
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007641<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007642
Mon P Wang28873102008-06-25 08:15:39 +00007643<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007644<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7645 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7646 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7647 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007648
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007649<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007650 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7651 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7652 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7653 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007654</pre>
7655
7656<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007657 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7658 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7659 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7660 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007661</pre>
7662
7663<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007664 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7665 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7666 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7667 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007668</pre>
7669
7670<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007671 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7672 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7673 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7674 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007675</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007676
Mon P Wang28873102008-06-25 08:15:39 +00007677<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007678<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007679 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7680 original value at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007681
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007682<h5>Arguments:</h5>
7683<p>These intrinsics take two arguments, the first a pointer to an integer value
7684 and the second an integer value. The result is also an integer value. These
7685 integer types can have any bit width, but they must all have the same bit
7686 width. The targets may only lower integer representations they support.</p>
7687
Mon P Wang28873102008-06-25 08:15:39 +00007688<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007689<p>These intrinsics does a series of operations atomically. They first load the
7690 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7691 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7692 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007693
7694<h5>Examples:</h5>
7695<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007696%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7697%ptr = bitcast i8* %mallocP to i32*
7698 store i32 7, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007699%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang28873102008-06-25 08:15:39 +00007700 <i>; yields {i32}:result0 = 7</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007701%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang28873102008-06-25 08:15:39 +00007702 <i>; yields {i32}:result1 = -2</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007703%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang28873102008-06-25 08:15:39 +00007704 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007705%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang28873102008-06-25 08:15:39 +00007706 <i>; yields {i32}:result3 = 8</i>
7707%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7708</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007709
Mon P Wang28873102008-06-25 08:15:39 +00007710</div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007711
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007712</div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007713
7714<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007715<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +00007716 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007717</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +00007718
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007719<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007720
7721<p>This class of intrinsics exists to information about the lifetime of memory
7722 objects and ranges where variables are immutable.</p>
7723
Nick Lewyckycc271862009-10-13 07:03:23 +00007724<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007725<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007726 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007727</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007728
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007729<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007730
7731<h5>Syntax:</h5>
7732<pre>
7733 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7734</pre>
7735
7736<h5>Overview:</h5>
7737<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7738 object's lifetime.</p>
7739
7740<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +00007741<p>The first argument is a constant integer representing the size of the
7742 object, or -1 if it is variable sized. The second argument is a pointer to
7743 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007744
7745<h5>Semantics:</h5>
7746<p>This intrinsic indicates that before this point in the code, the value of the
7747 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
Nick Lewycky8d336592009-10-27 16:56:58 +00007748 never be used and has an undefined value. A load from the pointer that
7749 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +00007750 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7751
7752</div>
7753
7754<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007755<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007756 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007757</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007758
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007759<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007760
7761<h5>Syntax:</h5>
7762<pre>
7763 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7764</pre>
7765
7766<h5>Overview:</h5>
7767<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7768 object's lifetime.</p>
7769
7770<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +00007771<p>The first argument is a constant integer representing the size of the
7772 object, or -1 if it is variable sized. The second argument is a pointer to
7773 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007774
7775<h5>Semantics:</h5>
7776<p>This intrinsic indicates that after this point in the code, the value of the
7777 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7778 never be used and has an undefined value. Any stores into the memory object
7779 following this intrinsic may be removed as dead.
7780
7781</div>
7782
7783<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007784<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007785 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007786</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007787
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007788<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007789
7790<h5>Syntax:</h5>
7791<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +00007792 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +00007793</pre>
7794
7795<h5>Overview:</h5>
7796<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7797 a memory object will not change.</p>
7798
7799<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +00007800<p>The first argument is a constant integer representing the size of the
7801 object, or -1 if it is variable sized. The second argument is a pointer to
7802 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007803
7804<h5>Semantics:</h5>
7805<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7806 the return value, the referenced memory location is constant and
7807 unchanging.</p>
7808
7809</div>
7810
7811<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007812<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007813 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007814</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007815
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007816<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007817
7818<h5>Syntax:</h5>
7819<pre>
7820 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7821</pre>
7822
7823<h5>Overview:</h5>
7824<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7825 a memory object are mutable.</p>
7826
7827<h5>Arguments:</h5>
7828<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +00007829 The second argument is a constant integer representing the size of the
7830 object, or -1 if it is variable sized and the third argument is a pointer
7831 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007832
7833<h5>Semantics:</h5>
7834<p>This intrinsic indicates that the memory is mutable again.</p>
7835
7836</div>
7837
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007838</div>
7839
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007840<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007841<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007842 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007843</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007844
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007845<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007846
7847<p>This class of intrinsics is designed to be generic and has no specific
7848 purpose.</p>
7849
Tanya Lattner6d806e92007-06-15 20:50:54 +00007850<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007851<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007852 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007853</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007854
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007855<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007856
7857<h5>Syntax:</h5>
7858<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007859 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 +00007860</pre>
7861
7862<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007863<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007864
7865<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007866<p>The first argument is a pointer to a value, the second is a pointer to a
7867 global string, the third is a pointer to a global string which is the source
7868 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007869
7870<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007871<p>This intrinsic allows annotation of local variables with arbitrary strings.
7872 This can be useful for special purpose optimizations that want to look for
7873 these annotations. These have no other defined use, they are ignored by code
7874 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007875
Tanya Lattner6d806e92007-06-15 20:50:54 +00007876</div>
7877
Tanya Lattnerb6367882007-09-21 22:59:12 +00007878<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007879<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +00007880 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007881</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007882
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007883<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007884
7885<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007886<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7887 any integer bit width.</p>
7888
Tanya Lattnerb6367882007-09-21 22:59:12 +00007889<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007890 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7891 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7892 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7893 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7894 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 +00007895</pre>
7896
7897<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007898<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007899
7900<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007901<p>The first argument is an integer value (result of some expression), the
7902 second is a pointer to a global string, the third is a pointer to a global
7903 string which is the source file name, and the last argument is the line
7904 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007905
7906<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007907<p>This intrinsic allows annotations to be put on arbitrary expressions with
7908 arbitrary strings. This can be useful for special purpose optimizations that
7909 want to look for these annotations. These have no other defined use, they
7910 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007911
Tanya Lattnerb6367882007-09-21 22:59:12 +00007912</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007913
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007914<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007915<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007916 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007917</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007918
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007919<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007920
7921<h5>Syntax:</h5>
7922<pre>
7923 declare void @llvm.trap()
7924</pre>
7925
7926<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007927<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007928
7929<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007930<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007931
7932<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007933<p>This intrinsics is lowered to the target dependent trap instruction. If the
7934 target does not have a trap instruction, this intrinsic will be lowered to
7935 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007936
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007937</div>
7938
Bill Wendling69e4adb2008-11-19 05:56:17 +00007939<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007940<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +00007941 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007942</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007943
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007944<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007945
Bill Wendling69e4adb2008-11-19 05:56:17 +00007946<h5>Syntax:</h5>
7947<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007948 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +00007949</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007950
Bill Wendling69e4adb2008-11-19 05:56:17 +00007951<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007952<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7953 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7954 ensure that it is placed on the stack before local variables.</p>
7955
Bill Wendling69e4adb2008-11-19 05:56:17 +00007956<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007957<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7958 arguments. The first argument is the value loaded from the stack
7959 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7960 that has enough space to hold the value of the guard.</p>
7961
Bill Wendling69e4adb2008-11-19 05:56:17 +00007962<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007963<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7964 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7965 stack. This is to ensure that if a local variable on the stack is
7966 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +00007967 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007968 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7969 function.</p>
7970
Bill Wendling69e4adb2008-11-19 05:56:17 +00007971</div>
7972
Eric Christopher0e671492009-11-30 08:03:53 +00007973<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007974<h4>
Eric Christopher0e671492009-11-30 08:03:53 +00007975 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007976</h4>
Eric Christopher0e671492009-11-30 08:03:53 +00007977
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007978<div>
Eric Christopher0e671492009-11-30 08:03:53 +00007979
7980<h5>Syntax:</h5>
7981<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007982 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
7983 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +00007984</pre>
7985
7986<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +00007987<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
7988 the optimizers to determine at compile time whether a) an operation (like
7989 memcpy) will overflow a buffer that corresponds to an object, or b) that a
7990 runtime check for overflow isn't necessary. An object in this context means
7991 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +00007992
7993<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +00007994<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +00007995 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +00007996 is a boolean 0 or 1. This argument determines whether you want the
7997 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher8295a0a2009-12-23 00:29:49 +00007998 1, variables are not allowed.</p>
7999
Eric Christopher0e671492009-11-30 08:03:53 +00008000<h5>Semantics:</h5>
8001<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +00008002 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8003 depending on the <tt>type</tt> argument, if the size cannot be determined at
8004 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +00008005
8006</div>
8007
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00008008</div>
8009
8010</div>
8011
Chris Lattner00950542001-06-06 20:29:01 +00008012<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00008013<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008014<address>
8015 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +00008016 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008017 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +00008018 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008019
8020 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +00008021 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008022 Last modified: $Date$
8023</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +00008024
Misha Brukman9d0919f2003-11-08 01:05:38 +00008025</body>
8026</html>