blob: 07720e51f2447c86199a86c6f4f367526ece51c2 [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>
Eli Friedman5b60e1b2011-07-20 21:35:53 +000056 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +000057 </ol>
58 </li>
Chris Lattner00950542001-06-06 20:29:01 +000059 <li><a href="#typesystem">Type System</a>
60 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +000061 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +000062 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +000063 <ol>
Nick Lewyckyec38da42009-09-27 00:45:11 +000064 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +000065 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen21fe99b2010-10-01 00:48:59 +000066 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +000067 <li><a href="#t_void">Void Type</a></li>
68 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky7a0370f2009-05-30 05:06:04 +000069 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000070 </ol>
71 </li>
Chris Lattner00950542001-06-06 20:29:01 +000072 <li><a href="#t_derived">Derived Types</a>
73 <ol>
Chris Lattnerfdfeb692010-02-12 20:49:41 +000074 <li><a href="#t_aggregate">Aggregate Types</a>
75 <ol>
76 <li><a href="#t_array">Array Type</a></li>
77 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattner1afcace2011-07-09 17:41:24 +000078 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattnerfdfeb692010-02-12 20:49:41 +000079 <li><a href="#t_vector">Vector Type</a></li>
80 </ol>
81 </li>
Misha Brukman9d0919f2003-11-08 01:05:38 +000082 <li><a href="#t_function">Function Type</a></li>
83 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +000084 </ol>
85 </li>
86 </ol>
87 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +000088 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +000089 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +000090 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner70882792009-02-28 18:32:25 +000091 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +000092 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
93 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanfff6c532010-04-22 23:14:21 +000094 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattnerf9d078e2009-10-27 21:19:13 +000095 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +000096 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattnerc3f59762004-12-09 17:30:23 +000097 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +000098 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +000099 <li><a href="#othervalues">Other Values</a>
100 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000101 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000102 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000103 </ol>
104 </li>
Chris Lattner857755c2009-07-20 05:55:19 +0000105 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
106 <ol>
107 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner401e10c2009-07-20 06:14:25 +0000108 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
109 Global Variable</a></li>
Chris Lattner857755c2009-07-20 05:55:19 +0000110 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
111 Global Variable</a></li>
112 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
113 Global Variable</a></li>
114 </ol>
115 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000116 <li><a href="#instref">Instruction Reference</a>
117 <ol>
118 <li><a href="#terminators">Terminator Instructions</a>
119 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000120 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
121 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000122 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerab21db72009-10-28 00:19:10 +0000123 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000124 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000125 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000126 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000127 </ol>
128 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000129 <li><a href="#binaryops">Binary Operations</a>
130 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000131 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000132 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000133 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000134 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000135 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000136 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000137 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
138 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
139 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000140 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
141 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
142 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000143 </ol>
144 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000145 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
146 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000147 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
148 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
149 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000150 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000151 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000152 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000153 </ol>
154 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000155 <li><a href="#vectorops">Vector Operations</a>
156 <ol>
157 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
158 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
159 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000160 </ol>
161 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000162 <li><a href="#aggregateops">Aggregate Operations</a>
163 <ol>
164 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
165 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
166 </ol>
167 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000168 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000169 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000170 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000171 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
172 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
173 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000174 </ol>
175 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000176 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000177 <ol>
178 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
179 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
180 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
181 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
182 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000183 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
184 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
185 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
186 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000187 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
188 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000189 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000190 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000191 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000192 <li><a href="#otherops">Other Operations</a>
193 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000194 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
195 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000196 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000197 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000198 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000199 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000200 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000201 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000202 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000203 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000204 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000205 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000206 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
207 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000208 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
209 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
210 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000211 </ol>
212 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000213 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
214 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000215 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
216 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
217 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000218 </ol>
219 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000220 <li><a href="#int_codegen">Code Generator Intrinsics</a>
221 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000222 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
223 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
224 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
225 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
226 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
227 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000228 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000229 </ol>
230 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000231 <li><a href="#int_libc">Standard C Library Intrinsics</a>
232 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000233 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
234 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
235 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000238 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
239 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohman08b280b2011-05-27 00:36:31 +0000241 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
242 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarich33390842011-07-08 21:39:21 +0000243 <li><a href="#int_fma">'<tt>llvm.fma.*</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
Dan Gohman129bd562011-06-16 16:03:13 +00001167 <dt><tt><b>nonlazybind</b></tt></dt>
1168 <dd>This attribute suppresses lazy symbol binding for the function. This
1169 may make calls to the function faster, at the cost of extra program
1170 startup time if the function is not called during program startup.</dd>
1171
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +00001172 <dt><tt><b>inlinehint</b></tt></dt>
1173 <dd>This attribute indicates that the source code contained a hint that inlining
1174 this function is desirable (such as the "inline" keyword in C/C++). It
1175 is just a hint; it imposes no requirements on the inliner.</dd>
1176
Nick Lewycky76ec37a2010-07-06 18:24:09 +00001177 <dt><tt><b>naked</b></tt></dt>
1178 <dd>This attribute disables prologue / epilogue emission for the function.
1179 This can have very system-specific consequences.</dd>
1180
1181 <dt><tt><b>noimplicitfloat</b></tt></dt>
1182 <dd>This attributes disables implicit floating point instructions.</dd>
1183
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001184 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001185 <dd>This attribute indicates that the inliner should never inline this
1186 function in any situation. This attribute may not be used together with
1187 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +00001188
Nick Lewycky76ec37a2010-07-06 18:24:09 +00001189 <dt><tt><b>noredzone</b></tt></dt>
1190 <dd>This attribute indicates that the code generator should not use a red
1191 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +00001192
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001193 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001194 <dd>This function attribute indicates that the function never returns
1195 normally. This produces undefined behavior at runtime if the function
1196 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +00001197
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001198 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001199 <dd>This function attribute indicates that the function never returns with an
1200 unwind or exceptional control flow. If the function does unwind, its
1201 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +00001202
Nick Lewycky76ec37a2010-07-06 18:24:09 +00001203 <dt><tt><b>optsize</b></tt></dt>
1204 <dd>This attribute suggests that optimization passes and code generator passes
1205 make choices that keep the code size of this function low, and otherwise
1206 do optimizations specifically to reduce code size.</dd>
1207
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001208 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001209 <dd>This attribute indicates that the function computes its result (or decides
1210 to unwind an exception) based strictly on its arguments, without
1211 dereferencing any pointer arguments or otherwise accessing any mutable
1212 state (e.g. memory, control registers, etc) visible to caller functions.
1213 It does not write through any pointer arguments
1214 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1215 changes any state visible to callers. This means that it cannot unwind
1216 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1217 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +00001218
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001219 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001220 <dd>This attribute indicates that the function does not write through any
1221 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1222 arguments) or otherwise modify any state (e.g. memory, control registers,
1223 etc) visible to caller functions. It may dereference pointer arguments
1224 and read state that may be set in the caller. A readonly function always
1225 returns the same value (or unwinds an exception identically) when called
1226 with the same set of arguments and global state. It cannot unwind an
1227 exception by calling the <tt>C++</tt> exception throwing methods, but may
1228 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +00001229
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001230 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001231 <dd>This attribute indicates that the function should emit a stack smashing
1232 protector. It is in the form of a "canary"&mdash;a random value placed on
1233 the stack before the local variables that's checked upon return from the
1234 function to see if it has been overwritten. A heuristic is used to
1235 determine if a function needs stack protectors or not.<br>
1236<br>
1237 If a function that has an <tt>ssp</tt> attribute is inlined into a
1238 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1239 function will have an <tt>ssp</tt> attribute.</dd>
1240
Bill Wendlingf82d40a2009-11-02 00:24:16 +00001241 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001242 <dd>This attribute indicates that the function should <em>always</em> emit a
1243 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001244 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1245<br>
1246 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1247 function that doesn't have an <tt>sspreq</tt> attribute or which has
1248 an <tt>ssp</tt> attribute, then the resulting function will have
1249 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +00001250</dl>
1251
Devang Patelf8b94812008-09-04 23:05:13 +00001252</div>
1253
1254<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001255<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +00001256 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001257</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001258
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001259<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001260
1261<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1262 the GCC "file scope inline asm" blocks. These blocks are internally
1263 concatenated by LLVM and treated as a single unit, but may be separated in
1264 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001265
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00001266<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00001267module asm "inline asm code goes here"
1268module asm "more can go here"
1269</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001270
1271<p>The strings can contain any character by escaping non-printable characters.
1272 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001273 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +00001274
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001275<p>The inline asm code is simply printed to the machine code .s file when
1276 assembly code is generated.</p>
1277
Chris Lattner4e9aba72006-01-23 23:23:47 +00001278</div>
Chris Lattnerfa730212004-12-09 16:11:40 +00001279
Reid Spencerde151942007-02-19 23:54:10 +00001280<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001281<h3>
Reid Spencerde151942007-02-19 23:54:10 +00001282 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001283</h3>
Reid Spencerde151942007-02-19 23:54:10 +00001284
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001285<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001286
Reid Spencerde151942007-02-19 23:54:10 +00001287<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001288 data is to be laid out in memory. The syntax for the data layout is
1289 simply:</p>
1290
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00001291<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001292target datalayout = "<i>layout specification</i>"
1293</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001294
1295<p>The <i>layout specification</i> consists of a list of specifications
1296 separated by the minus sign character ('-'). Each specification starts with
1297 a letter and may include other information after the letter to define some
1298 aspect of the data layout. The specifications accepted are as follows:</p>
1299
Reid Spencerde151942007-02-19 23:54:10 +00001300<dl>
1301 <dt><tt>E</tt></dt>
1302 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001303 bits with the most significance have the lowest address location.</dd>
1304
Reid Spencerde151942007-02-19 23:54:10 +00001305 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +00001306 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001307 the bits with the least significance have the lowest address
1308 location.</dd>
1309
Reid Spencerde151942007-02-19 23:54:10 +00001310 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001311 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001312 <i>preferred</i> alignments. All sizes are in bits. Specifying
1313 the <i>pref</i> alignment is optional. If omitted, the
1314 preceding <tt>:</tt> should be omitted too.</dd>
1315
Reid Spencerde151942007-02-19 23:54:10 +00001316 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1317 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001318 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1319
Reid Spencerde151942007-02-19 23:54:10 +00001320 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001321 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001322 <i>size</i>.</dd>
1323
Reid Spencerde151942007-02-19 23:54:10 +00001324 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001325 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +00001326 <i>size</i>. Only values of <i>size</i> that are supported by the target
1327 will work. 32 (float) and 64 (double) are supported on all targets;
1328 80 or 128 (different flavors of long double) are also supported on some
1329 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001330
Reid Spencerde151942007-02-19 23:54:10 +00001331 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1332 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001333 <i>size</i>.</dd>
1334
Daniel Dunbar87bde0b2009-06-08 22:17:53 +00001335 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1336 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001337 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +00001338
1339 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1340 <dd>This specifies a set of native integer widths for the target CPU
1341 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1342 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001343 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +00001344 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +00001345</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001346
Reid Spencerde151942007-02-19 23:54:10 +00001347<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +00001348 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001349 specifications in the <tt>datalayout</tt> keyword. The default specifications
1350 are given in this list:</p>
1351
Reid Spencerde151942007-02-19 23:54:10 +00001352<ul>
1353 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +00001354 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
Reid Spencerde151942007-02-19 23:54:10 +00001355 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1356 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1357 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1358 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +00001359 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencerde151942007-02-19 23:54:10 +00001360 alignment of 64-bits</li>
1361 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1362 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1363 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1364 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1365 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +00001366 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +00001367</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001368
1369<p>When LLVM is determining the alignment for a given type, it uses the
1370 following rules:</p>
1371
Reid Spencerde151942007-02-19 23:54:10 +00001372<ol>
1373 <li>If the type sought is an exact match for one of the specifications, that
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001374 specification is used.</li>
1375
Reid Spencerde151942007-02-19 23:54:10 +00001376 <li>If no match is found, and the type sought is an integer type, then the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001377 smallest integer type that is larger than the bitwidth of the sought type
1378 is used. If none of the specifications are larger than the bitwidth then
1379 the the largest integer type is used. For example, given the default
1380 specifications above, the i7 type will use the alignment of i8 (next
1381 largest) while both i65 and i256 will use the alignment of i64 (largest
1382 specified).</li>
1383
Reid Spencerde151942007-02-19 23:54:10 +00001384 <li>If no match is found, and the type sought is a vector type, then the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001385 largest vector type that is smaller than the sought vector type will be
1386 used as a fall back. This happens because &lt;128 x double&gt; can be
1387 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +00001388</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001389
Reid Spencerde151942007-02-19 23:54:10 +00001390</div>
Chris Lattnerfa730212004-12-09 16:11:40 +00001391
Dan Gohman556ca272009-07-27 18:07:55 +00001392<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001393<h3>
Dan Gohman556ca272009-07-27 18:07:55 +00001394 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001395</h3>
Dan Gohman556ca272009-07-27 18:07:55 +00001396
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001397<div>
Dan Gohman556ca272009-07-27 18:07:55 +00001398
Andreas Bolka55e459a2009-07-29 00:02:05 +00001399<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +00001400with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +00001401is undefined. Pointer values are associated with address ranges
1402according to the following rules:</p>
1403
1404<ul>
Dan Gohman1e109622010-07-02 18:41:32 +00001405 <li>A pointer value is associated with the addresses associated with
1406 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +00001407 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +00001408 range of the variable's storage.</li>
1409 <li>The result value of an allocation instruction is associated with
1410 the address range of the allocated storage.</li>
1411 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +00001412 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +00001413 <li>An integer constant other than zero or a pointer value returned
1414 from a function not defined within LLVM may be associated with address
1415 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +00001416 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +00001417 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +00001418</ul>
1419
1420<p>A pointer value is <i>based</i> on another pointer value according
1421 to the following rules:</p>
1422
1423<ul>
1424 <li>A pointer value formed from a
1425 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1426 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1427 <li>The result value of a
1428 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1429 of the <tt>bitcast</tt>.</li>
1430 <li>A pointer value formed by an
1431 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1432 pointer values that contribute (directly or indirectly) to the
1433 computation of the pointer's value.</li>
1434 <li>The "<i>based</i> on" relationship is transitive.</li>
1435</ul>
1436
1437<p>Note that this definition of <i>"based"</i> is intentionally
1438 similar to the definition of <i>"based"</i> in C99, though it is
1439 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +00001440
1441<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +00001442<tt><a href="#i_load">load</a></tt> merely indicates the size and
1443alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +00001444interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +00001445<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1446and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +00001447
1448<p>Consequently, type-based alias analysis, aka TBAA, aka
1449<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1450LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1451additional information which specialized optimization passes may use
1452to implement type-based alias analysis.</p>
1453
1454</div>
1455
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001456<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001457<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001458 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001459</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001460
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001461<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00001462
1463<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1464href="#i_store"><tt>store</tt></a>s, and <a
1465href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1466The optimizers must not change the number of volatile operations or change their
1467order of execution relative to other volatile operations. The optimizers
1468<i>may</i> change the order of volatile operations relative to non-volatile
1469operations. This is not Java's "volatile" and has no cross-thread
1470synchronization behavior.</p>
1471
1472</div>
1473
Eli Friedman5b60e1b2011-07-20 21:35:53 +00001474<!-- ======================================================================= -->
1475<h3>
1476 <a name="memmodel">Memory Model for Concurrent Operations</a>
1477</h3>
1478
1479<div>
1480
1481<p>The LLVM IR does not define any way to start parallel threads of execution
1482or to register signal handlers. Nonetheless, there are platform-specific
1483ways to create them, and we define LLVM IR's behavior in their presence. This
1484model is inspired by the C++0x memory model.</p>
1485
1486<p>We define a <i>happens-before</i> partial order as the least partial order
1487that</p>
1488<ul>
1489 <li>Is a superset of single-thread program order, and</li>
1490 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1491 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1492 by platform-specific techniques, like pthread locks, thread
1493 creation, thread joining, etc., and by the atomic operations described
1494 in the <a href="#int_atomics">Atomic intrinsics</a> section.</li>
1495</ul>
1496
1497<p>Note that program order does not introduce <i>happens-before</i> edges
1498between a thread and signals executing inside that thread.</p>
1499
1500<p>Every (defined) read operation (load instructions, memcpy, atomic
1501loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1502(defined) write operations (store instructions, atomic
1503stores/read-modify-writes, memcpy, etc.). For each byte, <var>R</var> reads the
1504value written by some write that it <i>may see</i>, given any relevant
1505<i>happens-before</i> constraints. <var>R<sub>byte</sub></var> may
1506see any write to the same byte, except:</p>
1507
1508<ul>
1509 <li>If <var>write<sub>1</sub></var> happens before
1510 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1511 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
1512 must not see <var>write<sub>1</sub></var>.
1513 <li>If <var>R<sub>byte</sub></var> happens before <var>write<sub>3</var>,
1514 then <var>R<sub>byte</sub></var> must not see
1515 <var>write<sub>3</sub></var>.
1516</ul>
1517
1518<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1519<ul>
1520 <li>If there is no write to the same byte that happens before
1521 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1522 <tt>undef</tt> for that byte.
1523 <li>If <var>R<sub>byte</sub></var> may see exactly one write,
1524 <var>R<sub>byte</sub></var> returns the value written by that
1525 write.</li>
1526 <li>If <var>R<sub>byte</sub></var> and all the writes it may see are
1527 atomic, it chooses one of those writes and returns it value.
1528 Given any two bytes in a given read <var>R</var>, if the set of
1529 writes <var>R<sub>byte</sub></var> may see is the same as the set
1530 of writes another byte may see, they will both choose the same write.
1531 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1532</ul>
1533
1534<p><var>R</var> returns the value composed of the series of bytes it read.
1535This implies that some bytes within the value may be <tt>undef</tt>
1536<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1537defines the semantics of the operation; it doesn't mean that targets will
1538emit more than one instruction to read the series of bytes.</p>
1539
1540<p>Note that in cases where none of the atomic intrinsics are used, this model
1541places only one restriction on IR transformations on top of what is required
1542for single-threaded execution: introducing a store to a byte which might not
1543otherwise be stored to can introduce undefined behavior.</p>
1544
1545<!-- FIXME: This model assumes all targets where concurrency is relevant have
1546a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1547none of the backends currently in the tree fall into this category; however,
1548there might be targets which care. If there are, we want a paragraph
1549like the following:
1550
1551Targets may specify that stores narrower than a certain width are not
1552available; on such a target, for the purposes of this model, treat any
1553non-atomic write with an alignment or width less than the minimum width
1554as if it writes to the relevant surrounding bytes.
1555-->
1556
1557</div>
1558
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001559</div>
1560
Chris Lattner00950542001-06-06 20:29:01 +00001561<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001562<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +00001563<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +00001564
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001565<div>
Chris Lattnerfa730212004-12-09 16:11:40 +00001566
Misha Brukman9d0919f2003-11-08 01:05:38 +00001567<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001568 intermediate representation. Being typed enables a number of optimizations
1569 to be performed on the intermediate representation directly, without having
1570 to do extra analyses on the side before the transformation. A strong type
1571 system makes it easier to read the generated code and enables novel analyses
1572 and transformations that are not feasible to perform on normal three address
1573 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +00001574
Chris Lattner00950542001-06-06 20:29:01 +00001575<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001576<h3>
1577 <a name="t_classifications">Type Classifications</a>
1578</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001579
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001580<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001581
1582<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001583
1584<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00001585 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001586 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +00001587 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001588 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +00001589 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +00001590 </tr>
1591 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001592 <td><a href="#t_floating">floating point</a></td>
1593 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +00001594 </tr>
1595 <tr>
1596 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +00001597 <td><a href="#t_integer">integer</a>,
1598 <a href="#t_floating">floating point</a>,
1599 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +00001600 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +00001601 <a href="#t_struct">structure</a>,
1602 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001603 <a href="#t_label">label</a>,
1604 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +00001605 </td>
Chris Lattner261efe92003-11-25 01:02:51 +00001606 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +00001607 <tr>
1608 <td><a href="#t_primitive">primitive</a></td>
1609 <td><a href="#t_label">label</a>,
1610 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +00001611 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001612 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001613 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001614 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +00001615 </tr>
1616 <tr>
1617 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001618 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +00001619 <a href="#t_function">function</a>,
1620 <a href="#t_pointer">pointer</a>,
1621 <a href="#t_struct">structure</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +00001622 <a href="#t_vector">vector</a>,
1623 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +00001624 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +00001625 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +00001626 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001627</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001628
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001629<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1630 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +00001631 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001632
Misha Brukman9d0919f2003-11-08 01:05:38 +00001633</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001634
Chris Lattner00950542001-06-06 20:29:01 +00001635<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001636<h3>
1637 <a name="t_primitive">Primitive Types</a>
1638</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +00001639
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001640<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001641
Chris Lattner4f69f462008-01-04 04:32:38 +00001642<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001643 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +00001644
1645<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001646<h4>
1647 <a name="t_integer">Integer Type</a>
1648</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +00001649
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001650<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +00001651
1652<h5>Overview:</h5>
1653<p>The integer type is a very simple type that simply specifies an arbitrary
1654 bit width for the integer type desired. Any bit width from 1 bit to
1655 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1656
1657<h5>Syntax:</h5>
1658<pre>
1659 iN
1660</pre>
1661
1662<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1663 value.</p>
1664
1665<h5>Examples:</h5>
1666<table class="layout">
1667 <tr class="layout">
1668 <td class="left"><tt>i1</tt></td>
1669 <td class="left">a single-bit integer.</td>
1670 </tr>
1671 <tr class="layout">
1672 <td class="left"><tt>i32</tt></td>
1673 <td class="left">a 32-bit integer.</td>
1674 </tr>
1675 <tr class="layout">
1676 <td class="left"><tt>i1942652</tt></td>
1677 <td class="left">a really big integer of over 1 million bits.</td>
1678 </tr>
1679</table>
1680
Nick Lewyckyec38da42009-09-27 00:45:11 +00001681</div>
1682
1683<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001684<h4>
1685 <a name="t_floating">Floating Point Types</a>
1686</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +00001687
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001688<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001689
1690<table>
1691 <tbody>
1692 <tr><th>Type</th><th>Description</th></tr>
1693 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1694 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1695 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1696 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1697 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1698 </tbody>
1699</table>
1700
Chris Lattner4f69f462008-01-04 04:32:38 +00001701</div>
1702
1703<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001704<h4>
1705 <a name="t_x86mmx">X86mmx Type</a>
1706</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001707
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001708<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001709
1710<h5>Overview:</h5>
1711<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>
1712
1713<h5>Syntax:</h5>
1714<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +00001715 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +00001716</pre>
1717
1718</div>
1719
1720<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001721<h4>
1722 <a name="t_void">Void Type</a>
1723</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +00001724
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001725<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001726
Chris Lattner4f69f462008-01-04 04:32:38 +00001727<h5>Overview:</h5>
1728<p>The void type does not represent any value and has no size.</p>
1729
1730<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +00001731<pre>
1732 void
1733</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001734
Chris Lattner4f69f462008-01-04 04:32:38 +00001735</div>
1736
1737<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001738<h4>
1739 <a name="t_label">Label Type</a>
1740</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +00001741
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001742<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001743
Chris Lattner4f69f462008-01-04 04:32:38 +00001744<h5>Overview:</h5>
1745<p>The label type represents code labels.</p>
1746
1747<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +00001748<pre>
1749 label
1750</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001751
Chris Lattner4f69f462008-01-04 04:32:38 +00001752</div>
1753
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001754<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001755<h4>
1756 <a name="t_metadata">Metadata Type</a>
1757</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001758
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001759<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001760
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001761<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +00001762<p>The metadata type represents embedded metadata. No derived types may be
1763 created from metadata except for <a href="#t_function">function</a>
1764 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001765
1766<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001767<pre>
1768 metadata
1769</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001770
Nick Lewycky7a0370f2009-05-30 05:06:04 +00001771</div>
1772
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001773</div>
Chris Lattner4f69f462008-01-04 04:32:38 +00001774
1775<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001776<h3>
1777 <a name="t_derived">Derived Types</a>
1778</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001779
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001780<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001781
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001782<p>The real power in LLVM comes from the derived types in the system. This is
1783 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +00001784 useful types. Each of these types contain one or more element types which
1785 may be a primitive type, or another derived type. For example, it is
1786 possible to have a two dimensional array, using an array as the element type
1787 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +00001788
Chris Lattner1afcace2011-07-09 17:41:24 +00001789</div>
1790
1791
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001792<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001793<h4>
1794 <a name="t_aggregate">Aggregate Types</a>
1795</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001796
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001797<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001798
1799<p>Aggregate Types are a subset of derived types that can contain multiple
1800 member types. <a href="#t_array">Arrays</a>,
Chris Lattner61c70e92010-08-28 04:09:24 +00001801 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1802 aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00001803
1804</div>
1805
Reid Spencer2b916312007-05-16 18:44:01 +00001806<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001807<h4>
1808 <a name="t_array">Array Type</a>
1809</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001810
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001811<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001812
Chris Lattner00950542001-06-06 20:29:01 +00001813<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00001814<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001815 sequentially in memory. The array type requires a size (number of elements)
1816 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001817
Chris Lattner7faa8832002-04-14 06:13:44 +00001818<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001819<pre>
1820 [&lt;# elements&gt; x &lt;elementtype&gt;]
1821</pre>
1822
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001823<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1824 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00001825
Chris Lattner7faa8832002-04-14 06:13:44 +00001826<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001827<table class="layout">
1828 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001829 <td class="left"><tt>[40 x i32]</tt></td>
1830 <td class="left">Array of 40 32-bit integer values.</td>
1831 </tr>
1832 <tr class="layout">
1833 <td class="left"><tt>[41 x i32]</tt></td>
1834 <td class="left">Array of 41 32-bit integer values.</td>
1835 </tr>
1836 <tr class="layout">
1837 <td class="left"><tt>[4 x i8]</tt></td>
1838 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001839 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001840</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001841<p>Here are some examples of multidimensional arrays:</p>
1842<table class="layout">
1843 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00001844 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1845 <td class="left">3x4 array of 32-bit integer values.</td>
1846 </tr>
1847 <tr class="layout">
1848 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1849 <td class="left">12x10 array of single precision floating point values.</td>
1850 </tr>
1851 <tr class="layout">
1852 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1853 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001854 </tr>
1855</table>
Chris Lattnere67a9512005-06-24 17:22:57 +00001856
Dan Gohman7657f6b2009-11-09 19:01:53 +00001857<p>There is no restriction on indexing beyond the end of the array implied by
1858 a static type (though there are restrictions on indexing beyond the bounds
1859 of an allocated object in some cases). This means that single-dimension
1860 'variable sized array' addressing can be implemented in LLVM with a zero
1861 length array type. An implementation of 'pascal style arrays' in LLVM could
1862 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +00001863
Misha Brukman9d0919f2003-11-08 01:05:38 +00001864</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001865
Chris Lattner00950542001-06-06 20:29:01 +00001866<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001867<h4>
1868 <a name="t_function">Function Type</a>
1869</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001870
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001871<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001872
Chris Lattner00950542001-06-06 20:29:01 +00001873<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001874<p>The function type can be thought of as a function signature. It consists of
1875 a return type and a list of formal parameter types. The return type of a
Chris Lattner61c70e92010-08-28 04:09:24 +00001876 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +00001877
Chris Lattner00950542001-06-06 20:29:01 +00001878<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001879<pre>
Nick Lewycky51386942009-09-27 07:55:32 +00001880 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001881</pre>
1882
John Criswell0ec250c2005-10-24 16:17:18 +00001883<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001884 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1885 which indicates that the function takes a variable number of arguments.
1886 Variable argument functions can access their arguments with
1887 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +00001888 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +00001889 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00001890
Chris Lattner00950542001-06-06 20:29:01 +00001891<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001892<table class="layout">
1893 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +00001894 <td class="left"><tt>i32 (i32)</tt></td>
1895 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001896 </td>
Reid Spencer92f82302006-12-31 07:18:34 +00001897 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +00001898 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +00001899 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001900 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +00001901 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1902 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +00001903 </td>
1904 </tr><tr class="layout">
1905 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00001906 <td class="left">A vararg function that takes at least one
1907 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1908 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +00001909 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +00001910 </td>
Devang Patela582f402008-03-24 05:35:41 +00001911 </tr><tr class="layout">
1912 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +00001913 <td class="left">A function taking an <tt>i32</tt>, returning a
1914 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +00001915 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001916 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001917</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00001918
Misha Brukman9d0919f2003-11-08 01:05:38 +00001919</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001920
Chris Lattner00950542001-06-06 20:29:01 +00001921<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001922<h4>
1923 <a name="t_struct">Structure Type</a>
1924</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001925
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001926<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001927
Chris Lattner00950542001-06-06 20:29:01 +00001928<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001929<p>The structure type is used to represent a collection of data members together
Chris Lattner1afcace2011-07-09 17:41:24 +00001930 in memory. The elements of a structure may be any type that has a size.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001931
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +00001932<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1933 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1934 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1935 Structures in registers are accessed using the
1936 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1937 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +00001938
1939<p>Structures may optionally be "packed" structures, which indicate that the
1940 alignment of the struct is one byte, and that there is no padding between
1941 the elements. In non-packed structs, padding between field types is defined
1942 by the target data string to match the underlying processor.</p>
1943
1944<p>Structures can either be "anonymous" or "named". An anonymous structure is
1945 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) and a named types
1946 are always defined at the top level with a name. Anonmyous types are uniqued
1947 by their contents and can never be recursive since there is no way to write
1948 one. Named types can be recursive.
1949</p>
1950
Chris Lattner00950542001-06-06 20:29:01 +00001951<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001952<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +00001953 %T1 = type { &lt;type list&gt; } <i>; Named normal struct type</i>
1954 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Named packed struct type</i>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001955</pre>
Chris Lattner1afcace2011-07-09 17:41:24 +00001956
Chris Lattner00950542001-06-06 20:29:01 +00001957<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001958<table class="layout">
1959 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001960 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1961 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattner1afcace2011-07-09 17:41:24 +00001962 </tr>
1963 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +00001964 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1965 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1966 second element is a <a href="#t_pointer">pointer</a> to a
1967 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1968 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00001969 </tr>
Chris Lattner1afcace2011-07-09 17:41:24 +00001970 <tr class="layout">
1971 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
1972 <td class="left">A packed struct known to be 5 bytes in size.</td>
1973 </tr>
Chris Lattner00950542001-06-06 20:29:01 +00001974</table>
Dan Gohmand8791e52009-01-24 15:58:40 +00001975
Misha Brukman9d0919f2003-11-08 01:05:38 +00001976</div>
Chris Lattner1afcace2011-07-09 17:41:24 +00001977
Chris Lattner00950542001-06-06 20:29:01 +00001978<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001979<h4>
Chris Lattner1afcace2011-07-09 17:41:24 +00001980 <a name="t_opaque">Opaque Type</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00001981</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001982
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00001983<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001984
Andrew Lenharth75e10682006-12-08 17:13:00 +00001985<h5>Overview:</h5>
Chris Lattner1afcace2011-07-09 17:41:24 +00001986<p>Opaque types are used to represent named structure types that do not have a
1987 body specified. This corresponds (for example) to the C notion of a forward
1988 declared structure.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001989
Andrew Lenharth75e10682006-12-08 17:13:00 +00001990<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001991<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +00001992 %X = type opaque
1993 %52 = type opaque
Bill Wendlingc39e3e02009-07-20 02:39:26 +00001994</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00001995
Andrew Lenharth75e10682006-12-08 17:13:00 +00001996<h5>Examples:</h5>
1997<table class="layout">
1998 <tr class="layout">
Chris Lattner1afcace2011-07-09 17:41:24 +00001999 <td class="left"><tt>opaque</tt></td>
2000 <td class="left">An opaque type.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +00002001 </tr>
2002</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002003
Andrew Lenharth75e10682006-12-08 17:13:00 +00002004</div>
2005
Chris Lattner1afcace2011-07-09 17:41:24 +00002006
2007
Andrew Lenharth75e10682006-12-08 17:13:00 +00002008<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002009<h4>
2010 <a name="t_pointer">Pointer Type</a>
2011</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +00002012
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002013<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002014
2015<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +00002016<p>The pointer type is used to specify memory locations.
2017 Pointers are commonly used to reference objects in memory.</p>
2018
2019<p>Pointer types may have an optional address space attribute defining the
2020 numbered address space where the pointed-to object resides. The default
2021 address space is number zero. The semantics of non-zero address
2022 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002023
2024<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2025 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +00002026
Chris Lattner7faa8832002-04-14 06:13:44 +00002027<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00002028<pre>
2029 &lt;type&gt; *
2030</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002031
Chris Lattner7faa8832002-04-14 06:13:44 +00002032<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00002033<table class="layout">
2034 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +00002035 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +00002036 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2037 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2038 </tr>
2039 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +00002040 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +00002041 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +00002042 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +00002043 <tt>i32</tt>.</td>
2044 </tr>
2045 <tr class="layout">
2046 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2047 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2048 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00002049 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +00002050</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002051
Misha Brukman9d0919f2003-11-08 01:05:38 +00002052</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +00002053
Chris Lattnera58561b2004-08-12 19:12:28 +00002054<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002055<h4>
2056 <a name="t_vector">Vector Type</a>
2057</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002058
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002059<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002060
Chris Lattnera58561b2004-08-12 19:12:28 +00002061<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002062<p>A vector type is a simple derived type that represents a vector of elements.
2063 Vector types are used when multiple primitive data are operated in parallel
2064 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +00002065 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002066 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002067
Chris Lattnera58561b2004-08-12 19:12:28 +00002068<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002069<pre>
2070 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2071</pre>
2072
Chris Lattner7d2e7be2010-10-10 18:20:35 +00002073<p>The number of elements is a constant integer value larger than 0; elementtype
2074 may be any integer or floating point type. Vectors of size zero are not
2075 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +00002076
Chris Lattnera58561b2004-08-12 19:12:28 +00002077<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +00002078<table class="layout">
2079 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +00002080 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2081 <td class="left">Vector of 4 32-bit integer values.</td>
2082 </tr>
2083 <tr class="layout">
2084 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2085 <td class="left">Vector of 8 32-bit floating-point values.</td>
2086 </tr>
2087 <tr class="layout">
2088 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2089 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +00002090 </tr>
2091</table>
Dan Gohmand8791e52009-01-24 15:58:40 +00002092
Misha Brukman9d0919f2003-11-08 01:05:38 +00002093</div>
2094
Chris Lattnerc3f59762004-12-09 17:30:23 +00002095<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002096<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002097<!-- *********************************************************************** -->
2098
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002099<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002100
2101<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002102 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002103
Chris Lattnerc3f59762004-12-09 17:30:23 +00002104<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002105<h3>
2106 <a name="simpleconstants">Simple Constants</a>
2107</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002108
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002109<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002110
2111<dl>
2112 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002113 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +00002114 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002115
2116 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002117 <dd>Standard integers (such as '4') are constants of
2118 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2119 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002120
2121 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002122 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002123 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2124 notation (see below). The assembler requires the exact decimal value of a
2125 floating-point constant. For example, the assembler accepts 1.25 but
2126 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2127 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002128
2129 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +00002130 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002131 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002132</dl>
2133
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002134<p>The one non-intuitive notation for constants is the hexadecimal form of
2135 floating point constants. For example, the form '<tt>double
2136 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2137 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2138 constants are required (and the only time that they are generated by the
2139 disassembler) is when a floating point constant must be emitted but it cannot
2140 be represented as a decimal floating point number in a reasonable number of
2141 digits. For example, NaN's, infinities, and other special values are
2142 represented in their IEEE hexadecimal format so that assembly and disassembly
2143 do not cause any bits to change in the constants.</p>
2144
Dale Johannesenbd5e5a82009-02-11 22:14:51 +00002145<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002146 represented using the 16-digit form shown above (which matches the IEEE754
2147 representation for double); float values must, however, be exactly
2148 representable as IEE754 single precision. Hexadecimal format is always used
2149 for long double, and there are three forms of long double. The 80-bit format
2150 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2151 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2152 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2153 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2154 currently supported target uses this format. Long doubles will only work if
2155 they match the long double format on your target. All hexadecimal formats
2156 are big-endian (sign bit at the left).</p>
2157
Dale Johannesen21fe99b2010-10-01 00:48:59 +00002158<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002159</div>
2160
2161<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002162<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +00002163<a name="aggregateconstants"></a> <!-- old anchor -->
2164<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002165</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002166
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002167<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002168
Chris Lattner70882792009-02-28 18:32:25 +00002169<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002170 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002171
2172<dl>
2173 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002174 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002175 type definitions (a comma separated list of elements, surrounded by braces
2176 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2177 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2178 Structure constants must have <a href="#t_struct">structure type</a>, and
2179 the number and types of elements must match those specified by the
2180 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002181
2182 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002183 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002184 definitions (a comma separated list of elements, surrounded by square
2185 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2186 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2187 the number and types of elements must match those specified by the
2188 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002189
Reid Spencer485bad12007-02-15 03:07:05 +00002190 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +00002191 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002192 definitions (a comma separated list of elements, surrounded by
2193 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2194 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2195 have <a href="#t_vector">vector type</a>, and the number and types of
2196 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002197
2198 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002199 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +00002200 value to zero of <em>any</em> type, including scalar and
2201 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002202 This is often used to avoid having to print large zero initializers
2203 (e.g. for large arrays) and is always exactly equivalent to using explicit
2204 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +00002205
2206 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +00002207 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002208 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2209 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2210 be interpreted as part of the instruction stream, metadata is a place to
2211 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002212</dl>
2213
2214</div>
2215
2216<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002217<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002218 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002219</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002220
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002221<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002222
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002223<p>The addresses of <a href="#globalvars">global variables</a>
2224 and <a href="#functionstructure">functions</a> are always implicitly valid
2225 (link-time) constants. These constants are explicitly referenced when
2226 the <a href="#identifiers">identifier for the global</a> is used and always
2227 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2228 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002229
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002230<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +00002231@X = global i32 17
2232@Y = global i32 42
2233@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +00002234</pre>
2235
2236</div>
2237
2238<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002239<h3>
2240 <a name="undefvalues">Undefined Values</a>
2241</h3>
2242
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002243<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002244
Chris Lattner48a109c2009-09-07 22:52:39 +00002245<p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
Benjamin Kramer8040cd32009-10-12 14:46:08 +00002246 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +00002247 Undefined values may be of any type (other than '<tt>label</tt>'
2248 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002249
Chris Lattnerc608cb12009-09-11 01:49:31 +00002250<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +00002251 program is well defined no matter what value is used. This gives the
2252 compiler more freedom to optimize. Here are some examples of (potentially
2253 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002254
Chris Lattner48a109c2009-09-07 22:52:39 +00002255
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002256<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002257 %A = add %X, undef
2258 %B = sub %X, undef
2259 %C = xor %X, undef
2260Safe:
2261 %A = undef
2262 %B = undef
2263 %C = undef
2264</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002265
2266<p>This is safe because all of the output bits are affected by the undef bits.
Bill Wendling1b383ba2010-10-27 01:07:41 +00002267 Any output bit can have a zero or one depending on the input bits.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +00002268
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002269<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002270 %A = or %X, undef
2271 %B = and %X, undef
2272Safe:
2273 %A = -1
2274 %B = 0
2275Unsafe:
2276 %A = undef
2277 %B = undef
2278</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002279
2280<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +00002281 For example, if <tt>%X</tt> has a zero bit, then the output of the
2282 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2283 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2284 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2285 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2286 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2287 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2288 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +00002289
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002290<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002291 %A = select undef, %X, %Y
2292 %B = select undef, 42, %Y
2293 %C = select %X, %Y, undef
2294Safe:
2295 %A = %X (or %Y)
2296 %B = 42 (or %Y)
2297 %C = %Y
2298Unsafe:
2299 %A = undef
2300 %B = undef
2301 %C = undef
2302</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002303
Bill Wendling1b383ba2010-10-27 01:07:41 +00002304<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2305 branch) conditions can go <em>either way</em>, but they have to come from one
2306 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2307 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2308 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2309 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2310 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2311 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +00002312
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002313<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +00002314 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002315
Chris Lattner48a109c2009-09-07 22:52:39 +00002316 %B = undef
2317 %C = xor %B, %B
2318
2319 %D = undef
2320 %E = icmp lt %D, 4
2321 %F = icmp gte %D, 4
2322
2323Safe:
2324 %A = undef
2325 %B = undef
2326 %C = undef
2327 %D = undef
2328 %E = undef
2329 %F = undef
2330</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +00002331
Bill Wendling1b383ba2010-10-27 01:07:41 +00002332<p>This example points out that two '<tt>undef</tt>' operands are not
2333 necessarily the same. This can be surprising to people (and also matches C
2334 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2335 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2336 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2337 its value over its "live range". This is true because the variable doesn't
2338 actually <em>have a live range</em>. Instead, the value is logically read
2339 from arbitrary registers that happen to be around when needed, so the value
2340 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2341 need to have the same semantics or the core LLVM "replace all uses with"
2342 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002343
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002344<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +00002345 %A = fdiv undef, %X
2346 %B = fdiv %X, undef
2347Safe:
2348 %A = undef
2349b: unreachable
2350</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002351
2352<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +00002353 value</em> and <em>undefined behavior</em>. An undefined value (like
2354 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2355 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2356 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2357 defined on SNaN's. However, in the second example, we can make a more
2358 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2359 arbitrary value, we are allowed to assume that it could be zero. Since a
2360 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2361 the operation does not execute at all. This allows us to delete the divide and
2362 all code after it. Because the undefined operation "can't happen", the
2363 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002364
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002365<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +00002366a: store undef -> %X
2367b: store %X -> undef
2368Safe:
2369a: &lt;deleted&gt;
2370b: unreachable
2371</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002372
Bill Wendling1b383ba2010-10-27 01:07:41 +00002373<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2374 undefined value can be assumed to not have any effect; we can assume that the
2375 value is overwritten with bits that happen to match what was already there.
2376 However, a store <em>to</em> an undefined location could clobber arbitrary
2377 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +00002378
Chris Lattnerc3f59762004-12-09 17:30:23 +00002379</div>
2380
2381<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002382<h3>
2383 <a name="trapvalues">Trap Values</a>
2384</h3>
2385
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002386<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002387
Dan Gohmanc68ce062010-04-26 20:21:21 +00002388<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanfff6c532010-04-22 23:14:21 +00002389 instead of representing an unspecified bit pattern, they represent the
2390 fact that an instruction or constant expression which cannot evoke side
2391 effects has nevertheless detected a condition which results in undefined
Dan Gohmanc68ce062010-04-26 20:21:21 +00002392 behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002393
Dan Gohman34b3d992010-04-28 00:49:41 +00002394<p>There is currently no way of representing a trap value in the IR; they
Dan Gohman855abed2010-05-03 14:51:43 +00002395 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +00002396 <a href="#i_add"><tt>add</tt></a> with the <tt>nsw</tt> flag.</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002397
Dan Gohman34b3d992010-04-28 00:49:41 +00002398<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002399
Dan Gohman34b3d992010-04-28 00:49:41 +00002400<ul>
2401<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2402 their operands.</li>
2403
2404<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2405 to their dynamic predecessor basic block.</li>
2406
2407<li>Function arguments depend on the corresponding actual argument values in
2408 the dynamic callers of their functions.</li>
2409
2410<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2411 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2412 control back to them.</li>
2413
Dan Gohmanb5328162010-05-03 14:55:22 +00002414<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2415 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2416 or exception-throwing call instructions that dynamically transfer control
2417 back to them.</li>
2418
Dan Gohman34b3d992010-04-28 00:49:41 +00002419<li>Non-volatile loads and stores depend on the most recent stores to all of the
2420 referenced memory addresses, following the order in the IR
2421 (including loads and stores implied by intrinsics such as
2422 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2423
Dan Gohman7c24ff12010-05-03 14:59:34 +00002424<!-- TODO: In the case of multiple threads, this only applies if the store
2425 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002426
Dan Gohman34b3d992010-04-28 00:49:41 +00002427<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002428
Dan Gohman34b3d992010-04-28 00:49:41 +00002429<li>An instruction with externally visible side effects depends on the most
2430 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +00002431 the order in the IR. (This includes
2432 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002433
Dan Gohmanb5328162010-05-03 14:55:22 +00002434<li>An instruction <i>control-depends</i> on a
2435 <a href="#terminators">terminator instruction</a>
2436 if the terminator instruction has multiple successors and the instruction
2437 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +00002438 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +00002439
Dan Gohmanca4cac42011-04-12 23:05:59 +00002440<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2441 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +00002442 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +00002443 successor.</li>
2444
Dan Gohman34b3d992010-04-28 00:49:41 +00002445<li>Dependence is transitive.</li>
2446
2447</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +00002448
2449<p>Whenever a trap value is generated, all values which depend on it evaluate
2450 to trap. If they have side effects, the evoke their side effects as if each
2451 operand with a trap value were undef. If they have externally-visible side
2452 effects, the behavior is undefined.</p>
2453
2454<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +00002455
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002456<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002457entry:
2458 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +00002459 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2460 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2461 store i32 0, i32* %trap_yet_again ; undefined behavior
2462
2463 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2464 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2465
2466 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2467
2468 %narrowaddr = bitcast i32* @g to i16*
2469 %wideaddr = bitcast i32* @g to i64*
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002470 %trap3 = load i16* %narrowaddr ; Returns a trap value.
2471 %trap4 = load i64* %wideaddr ; Returns a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +00002472
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002473 %cmp = icmp slt i32 %trap, 0 ; Returns a trap value.
2474 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002475
2476true:
Dan Gohman34b3d992010-04-28 00:49:41 +00002477 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2478 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002479 br label %end
2480
2481end:
2482 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2483 ; Both edges into this PHI are
2484 ; control-dependent on %cmp, so this
Dan Gohman34b3d992010-04-28 00:49:41 +00002485 ; always results in a trap value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002486
Dan Gohmanca4cac42011-04-12 23:05:59 +00002487 volatile store i32 0, i32* @g ; This would depend on the store in %true
2488 ; if %cmp is true, or the store in %entry
2489 ; otherwise, so this is undefined behavior.
2490
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002491 br i1 %cmp, label %second_true, label %second_end
Dan Gohmanca4cac42011-04-12 23:05:59 +00002492 ; The same branch again, but this time the
2493 ; true block doesn't have side effects.
2494
2495second_true:
2496 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002497 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +00002498
2499second_end:
2500 volatile store i32 0, i32* @g ; This time, the instruction always depends
2501 ; on the store in %end. Also, it is
2502 ; control-equivalent to %end, so this is
Nick Lewycky64f9fb12011-05-16 19:29:30 +00002503 ; well-defined (again, ignoring earlier
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002504 ; undefined behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +00002505</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +00002506
Dan Gohmanfff6c532010-04-22 23:14:21 +00002507</div>
2508
2509<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002510<h3>
2511 <a name="blockaddress">Addresses of Basic Blocks</a>
2512</h3>
2513
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002514<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +00002515
Chris Lattnercdfc9402009-11-01 01:27:45 +00002516<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +00002517
2518<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +00002519 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +00002520 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002521
Chris Lattnerc6f44362009-10-27 21:01:34 +00002522<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +00002523 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2524 comparisons against null. Pointer equality tests between labels addresses
2525 results in undefined behavior &mdash; though, again, comparison against null
2526 is ok, and no label is equal to the null pointer. This may be passed around
2527 as an opaque pointer sized value as long as the bits are not inspected. This
2528 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2529 long as the original value is reconstituted before the <tt>indirectbr</tt>
2530 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00002531
Bill Wendling1b383ba2010-10-27 01:07:41 +00002532<p>Finally, some targets may provide defined semantics when using the value as
2533 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +00002534
2535</div>
2536
2537
2538<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002539<h3>
2540 <a name="constantexprs">Constant Expressions</a>
2541</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002542
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002543<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002544
2545<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002546 to be used as constants. Constant expressions may be of
2547 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2548 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +00002549 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002550
2551<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002552 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002553 <dd>Truncate a constant to another type. The bit size of CST must be larger
2554 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002555
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002556 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002557 <dd>Zero extend a constant to another type. The bit size of CST must be
Duncan Sands28afd432010-07-13 12:06:14 +00002558 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002559
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002560 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002561 <dd>Sign extend a constant to another type. The bit size of CST must be
Duncan Sands28afd432010-07-13 12:06:14 +00002562 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002563
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002564 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002565 <dd>Truncate a floating point constant to another floating point type. The
2566 size of CST must be larger than the size of TYPE. Both types must be
2567 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002568
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002569 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002570 <dd>Floating point extend a constant to another type. The size of CST must be
2571 smaller or equal to the size of TYPE. Both types must be floating
2572 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002573
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002574 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002575 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002576 constant. TYPE must be a scalar or vector integer type. CST must be of
2577 scalar or vector floating point type. Both CST and TYPE must be scalars,
2578 or vectors of the same number of elements. If the value won't fit in the
2579 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002580
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002581 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002582 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002583 constant. TYPE must be a scalar or vector integer type. CST must be of
2584 scalar or vector floating point type. Both CST and TYPE must be scalars,
2585 or vectors of the same number of elements. If the value won't fit in the
2586 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002587
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002588 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002589 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002590 constant. TYPE must be a scalar or vector floating point type. CST must be
2591 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2592 vectors of the same number of elements. If the value won't fit in the
2593 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002594
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002595 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002596 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002597 constant. TYPE must be a scalar or vector floating point type. CST must be
2598 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2599 vectors of the same number of elements. If the value won't fit in the
2600 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00002601
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002602 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +00002603 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002604 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2605 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2606 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +00002607
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002608 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002609 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2610 type. CST must be of integer type. The CST value is zero extended,
2611 truncated, or unchanged to make it fit in a pointer size. This one is
2612 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +00002613
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002614 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +00002615 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2616 are the same as those for the <a href="#i_bitcast">bitcast
2617 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002618
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002619 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2620 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002621 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002622 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2623 instruction, the index list may have zero or more indexes, which are
2624 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002625
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002626 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002627 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +00002628
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002629 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +00002630 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2631
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002632 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +00002633 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +00002634
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002635 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002636 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2637 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +00002638
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002639 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002640 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2641 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +00002642
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002643 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002644 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2645 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +00002646
Nick Lewycky9e130ce2010-05-29 06:44:15 +00002647 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2648 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2649 constants. The index list is interpreted in a similar manner as indices in
2650 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2651 index value must be specified.</dd>
2652
2653 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2654 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2655 constants. The index list is interpreted in a similar manner as indices in
2656 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2657 index value must be specified.</dd>
2658
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00002659 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002660 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2661 be any of the <a href="#binaryops">binary</a>
2662 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2663 on operands are the same as those for the corresponding instruction
2664 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002665</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002666
Chris Lattnerc3f59762004-12-09 17:30:23 +00002667</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +00002668
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002669</div>
2670
Chris Lattner00950542001-06-06 20:29:01 +00002671<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002672<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +00002673<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002674<div>
Chris Lattnere87d6532006-01-25 23:47:57 +00002675<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002676<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +00002677<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002678</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +00002679
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002680<div>
Chris Lattnere87d6532006-01-25 23:47:57 +00002681
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002682<p>LLVM supports inline assembler expressions (as opposed
2683 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2684 a special value. This value represents the inline assembler as a string
2685 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +00002686 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +00002687 expression has side effects, and a flag indicating whether the function
2688 containing the asm needs to align its stack conservatively. An example
2689 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002690
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002691<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00002692i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +00002693</pre>
2694
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002695<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2696 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2697 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002698
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002699<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00002700%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattnere87d6532006-01-25 23:47:57 +00002701</pre>
2702
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002703<p>Inline asms with side effects not visible in the constraint list must be
2704 marked as having side effects. This is done through the use of the
2705 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002706
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002707<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00002708call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +00002709</pre>
2710
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +00002711<p>In some cases inline asms will contain code that will not work unless the
2712 stack is aligned in some way, such as calls or SSE instructions on x86,
2713 yet will not contain code that does that alignment within the asm.
2714 The compiler should make conservative assumptions about what the asm might
2715 contain and should generate its usual stack alignment code in the prologue
2716 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +00002717
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002718<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +00002719call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +00002720</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +00002721
2722<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2723 first.</p>
2724
Chris Lattnere87d6532006-01-25 23:47:57 +00002725<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002726 documented here. Constraints on what can be done (e.g. duplication, moving,
2727 etc need to be documented). This is probably best done by reference to
2728 another document that covers inline asm from a holistic perspective.</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002729
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002730<h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002731<a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002732</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002733
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002734<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002735
2736<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattnerce1b9ad2010-11-17 08:20:42 +00002737 attached to it that contains a list of constant integers. If present, the
2738 code generator will use the integer as the location cookie value when report
Chris Lattnercf9a4152010-04-07 05:38:05 +00002739 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman1c70c002010-04-28 00:36:01 +00002740 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattnercf9a4152010-04-07 05:38:05 +00002741 source code that produced it. For example:</p>
2742
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002743<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +00002744call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2745...
2746!42 = !{ i32 1234567 }
2747</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +00002748
2749<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 +00002750 IR. If the MDNode contains multiple constants, the code generator will use
2751 the one that corresponds to the line of the asm that the error occurs on.</p>
Chris Lattnere87d6532006-01-25 23:47:57 +00002752
2753</div>
2754
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002755</div>
2756
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002757<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002758<h3>
2759 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2760</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002761
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002762<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002763
2764<p>LLVM IR allows metadata to be attached to instructions in the program that
2765 can convey extra information about the code to the optimizers and code
2766 generator. One example application of metadata is source-level debug
2767 information. There are two metadata primitives: strings and nodes. All
2768 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2769 preceding exclamation point ('<tt>!</tt>').</p>
2770
2771<p>A metadata string is a string surrounded by double quotes. It can contain
2772 any character by escaping non-printable characters with "\xx" where "xx" is
2773 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2774
2775<p>Metadata nodes are represented with notation similar to structure constants
2776 (a comma separated list of elements, surrounded by braces and preceded by an
2777 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2778 10}</tt>". Metadata nodes can have any values as their operand.</p>
2779
2780<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2781 metadata nodes, which can be looked up in the module symbol table. For
2782 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2783
Devang Patele1d50cd2010-03-04 23:44:48 +00002784<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002785 function is using two metadata arguments.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +00002786
Bill Wendling9ff5de92011-03-02 02:17:11 +00002787<div class="doc_code">
2788<pre>
2789call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2790</pre>
2791</div>
Devang Patele1d50cd2010-03-04 23:44:48 +00002792
2793<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00002794 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +00002795
Bill Wendling9ff5de92011-03-02 02:17:11 +00002796<div class="doc_code">
2797<pre>
2798%indvar.next = add i64 %indvar, 1, !dbg !21
2799</pre>
2800</div>
2801
Chris Lattnere6a5ddd2010-01-15 21:50:19 +00002802</div>
2803
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002804</div>
Chris Lattner857755c2009-07-20 05:55:19 +00002805
2806<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002807<h2>
Chris Lattner857755c2009-07-20 05:55:19 +00002808 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002809</h2>
Chris Lattner857755c2009-07-20 05:55:19 +00002810<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002811<div>
Chris Lattner857755c2009-07-20 05:55:19 +00002812<p>LLVM has a number of "magic" global variables that contain data that affect
2813code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +00002814of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2815section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2816by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +00002817
2818<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002819<h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002820<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002821</h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002822
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002823<div>
Chris Lattner857755c2009-07-20 05:55:19 +00002824
2825<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2826href="#linkage_appending">appending linkage</a>. This array contains a list of
2827pointers to global variables and functions which may optionally have a pointer
2828cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2829
2830<pre>
2831 @X = global i8 4
2832 @Y = global i32 123
2833
2834 @llvm.used = appending global [2 x i8*] [
2835 i8* @X,
2836 i8* bitcast (i32* @Y to i8*)
2837 ], section "llvm.metadata"
2838</pre>
2839
2840<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2841compiler, assembler, and linker are required to treat the symbol as if there is
2842a reference to the global that it cannot see. For example, if a variable has
2843internal linkage and no references other than that from the <tt>@llvm.used</tt>
2844list, it cannot be deleted. This is commonly used to represent references from
2845inline asms and other things the compiler cannot "see", and corresponds to
2846"attribute((used))" in GNU C.</p>
2847
2848<p>On some targets, the code generator must emit a directive to the assembler or
2849object file to prevent the assembler and linker from molesting the symbol.</p>
2850
2851</div>
2852
2853<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002854<h3>
2855 <a name="intg_compiler_used">
2856 The '<tt>llvm.compiler.used</tt>' Global Variable
2857 </a>
2858</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +00002859
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002860<div>
Chris Lattner401e10c2009-07-20 06:14:25 +00002861
2862<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2863<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2864touching the symbol. On targets that support it, this allows an intelligent
2865linker to optimize references to the symbol without being impeded as it would be
2866by <tt>@llvm.used</tt>.</p>
2867
2868<p>This is a rare construct that should only be used in rare circumstances, and
2869should not be exposed to source languages.</p>
2870
2871</div>
2872
2873<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002874<h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002875<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002876</h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002877
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002878<div>
David Chisnalle31e9962010-04-30 19:23:49 +00002879<pre>
2880%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +00002881@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +00002882</pre>
2883<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.
2884</p>
Chris Lattner857755c2009-07-20 05:55:19 +00002885
2886</div>
2887
2888<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002889<h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002890<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002891</h3>
Chris Lattner857755c2009-07-20 05:55:19 +00002892
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002893<div>
David Chisnalle31e9962010-04-30 19:23:49 +00002894<pre>
2895%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +00002896@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +00002897</pre>
Chris Lattner857755c2009-07-20 05:55:19 +00002898
David Chisnalle31e9962010-04-30 19:23:49 +00002899<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.
2900</p>
Chris Lattner857755c2009-07-20 05:55:19 +00002901
2902</div>
2903
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002904</div>
Chris Lattner857755c2009-07-20 05:55:19 +00002905
Chris Lattnere87d6532006-01-25 23:47:57 +00002906<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002907<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +00002908<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +00002909
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002910<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002911
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002912<p>The LLVM instruction set consists of several different classifications of
2913 instructions: <a href="#terminators">terminator
2914 instructions</a>, <a href="#binaryops">binary instructions</a>,
2915 <a href="#bitwiseops">bitwise binary instructions</a>,
2916 <a href="#memoryops">memory instructions</a>, and
2917 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002918
Chris Lattner00950542001-06-06 20:29:01 +00002919<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002920<h3>
2921 <a name="terminators">Terminator Instructions</a>
2922</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002923
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002924<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002925
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002926<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2927 in a program ends with a "Terminator" instruction, which indicates which
2928 block should be executed after the current block is finished. These
2929 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2930 control flow, not values (the one exception being the
2931 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2932
Duncan Sands83821c82010-04-15 20:35:54 +00002933<p>There are seven different terminator instructions: the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002934 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2935 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2936 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling21c346e2009-11-02 00:25:26 +00002937 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002938 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2939 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2940 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +00002941
Chris Lattner00950542001-06-06 20:29:01 +00002942<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002943<h4>
2944 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
2945</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002946
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002947<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002948
Chris Lattner00950542001-06-06 20:29:01 +00002949<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +00002950<pre>
2951 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002952 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +00002953</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002954
Chris Lattner00950542001-06-06 20:29:01 +00002955<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002956<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2957 a value) from a function back to the caller.</p>
2958
2959<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2960 value and then causes control flow, and one that just causes control flow to
2961 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002962
Chris Lattner00950542001-06-06 20:29:01 +00002963<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002964<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2965 return value. The type of the return value must be a
2966 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +00002967
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002968<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2969 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2970 value or a return value with a type that does not match its type, or if it
2971 has a void return type and contains a '<tt>ret</tt>' instruction with a
2972 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002973
Chris Lattner00950542001-06-06 20:29:01 +00002974<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002975<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2976 the calling function's context. If the caller is a
2977 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2978 instruction after the call. If the caller was an
2979 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2980 the beginning of the "normal" destination block. If the instruction returns
2981 a value, that value shall set the call or invoke instruction's return
2982 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002983
Chris Lattner00950542001-06-06 20:29:01 +00002984<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +00002985<pre>
2986 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +00002987 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +00002988 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Chris Lattner00950542001-06-06 20:29:01 +00002989</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +00002990
Misha Brukman9d0919f2003-11-08 01:05:38 +00002991</div>
Chris Lattner00950542001-06-06 20:29:01 +00002992<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00002993<h4>
2994 <a name="i_br">'<tt>br</tt>' Instruction</a>
2995</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002996
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00002997<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00002998
Chris Lattner00950542001-06-06 20:29:01 +00002999<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003000<pre>
3001 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 +00003002</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003003
Chris Lattner00950542001-06-06 20:29:01 +00003004<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003005<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3006 different basic block in the current function. There are two forms of this
3007 instruction, corresponding to a conditional branch and an unconditional
3008 branch.</p>
3009
Chris Lattner00950542001-06-06 20:29:01 +00003010<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003011<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3012 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3013 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3014 target.</p>
3015
Chris Lattner00950542001-06-06 20:29:01 +00003016<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +00003017<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003018 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3019 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3020 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3021
Chris Lattner00950542001-06-06 20:29:01 +00003022<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +00003023<pre>
3024Test:
3025 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3026 br i1 %cond, label %IfEqual, label %IfUnequal
3027IfEqual:
3028 <a href="#i_ret">ret</a> i32 1
3029IfUnequal:
3030 <a href="#i_ret">ret</a> i32 0
3031</pre>
3032
Misha Brukman9d0919f2003-11-08 01:05:38 +00003033</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003034
Chris Lattner00950542001-06-06 20:29:01 +00003035<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003036<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003037 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003038</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003039
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003040<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003041
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003042<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003043<pre>
3044 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3045</pre>
3046
Chris Lattner00950542001-06-06 20:29:01 +00003047<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003048<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003049 several different places. It is a generalization of the '<tt>br</tt>'
3050 instruction, allowing a branch to occur to one of many possible
3051 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003052
Chris Lattner00950542001-06-06 20:29:01 +00003053<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003054<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003055 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3056 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3057 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003058
Chris Lattner00950542001-06-06 20:29:01 +00003059<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +00003060<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003061 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3062 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +00003063 transferred to the corresponding destination; otherwise, control flow is
3064 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +00003065
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003066<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003067<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003068 <tt>switch</tt> instruction, this instruction may be code generated in
3069 different ways. For example, it could be generated as a series of chained
3070 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003071
3072<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003073<pre>
3074 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00003075 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +00003076 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003077
3078 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003079 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +00003080
3081 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +00003082 switch i32 %val, label %otherwise [ i32 0, label %onzero
3083 i32 1, label %onone
3084 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +00003085</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003086
Misha Brukman9d0919f2003-11-08 01:05:38 +00003087</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003088
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003089
3090<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003091<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +00003092 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003093</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003094
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003095<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003096
3097<h5>Syntax:</h5>
3098<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +00003099 indirectbr &lt;somety&gt;* &lt;address&gt;, [ label &lt;dest1&gt;, label &lt;dest2&gt;, ... ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003100</pre>
3101
3102<h5>Overview:</h5>
3103
Chris Lattnerab21db72009-10-28 00:19:10 +00003104<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003105 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +00003106 "<tt>address</tt>". Address must be derived from a <a
3107 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003108
3109<h5>Arguments:</h5>
3110
3111<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3112 rest of the arguments indicate the full set of possible destinations that the
3113 address may point to. Blocks are allowed to occur multiple times in the
3114 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003115
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003116<p>This destination list is required so that dataflow analysis has an accurate
3117 understanding of the CFG.</p>
3118
3119<h5>Semantics:</h5>
3120
3121<p>Control transfers to the block specified in the address argument. All
3122 possible destination blocks must be listed in the label list, otherwise this
3123 instruction has undefined behavior. This implies that jumps to labels
3124 defined in other functions have undefined behavior as well.</p>
3125
3126<h5>Implementation:</h5>
3127
3128<p>This is typically implemented with a jump through a register.</p>
3129
3130<h5>Example:</h5>
3131<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +00003132 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +00003133</pre>
3134
3135</div>
3136
3137
Chris Lattner00950542001-06-06 20:29:01 +00003138<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003139<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003140 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003141</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003142
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003143<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003144
Chris Lattner00950542001-06-06 20:29:01 +00003145<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003146<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +00003147 &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 +00003148 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003149</pre>
3150
Chris Lattner6536cfe2002-05-06 22:08:29 +00003151<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003152<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003153 function, with the possibility of control flow transfer to either the
3154 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3155 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3156 control flow will return to the "normal" label. If the callee (or any
3157 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3158 instruction, control is interrupted and continued at the dynamically nearest
3159 "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003160
Chris Lattner00950542001-06-06 20:29:01 +00003161<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003162<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003163
Chris Lattner00950542001-06-06 20:29:01 +00003164<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003165 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3166 convention</a> the call should use. If none is specified, the call
3167 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +00003168
3169 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003170 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3171 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +00003172
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003173 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003174 function value being invoked. In most cases, this is a direct function
3175 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3176 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003177
3178 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003179 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003180
3181 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +00003182 signature argument types and parameter attributes. All arguments must be
3183 of <a href="#t_firstclass">first class</a> type. If the function
3184 signature indicates the function accepts a variable number of arguments,
3185 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003186
3187 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003188 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003189
3190 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003191 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003192
Devang Patel307e8ab2008-10-07 17:48:33 +00003193 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003194 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3195 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +00003196</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003197
Chris Lattner00950542001-06-06 20:29:01 +00003198<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003199<p>This instruction is designed to operate as a standard
3200 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3201 primary difference is that it establishes an association with a label, which
3202 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003203
3204<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003205 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3206 exception. Additionally, this is important for implementation of
3207 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003208
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003209<p>For the purposes of the SSA form, the definition of the value returned by the
3210 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3211 block to the "normal" label. If the callee unwinds then no return value is
3212 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +00003213
Chris Lattnerdf7a6802010-01-15 18:08:37 +00003214<p>Note that the code generator does not yet completely support unwind, and
3215that the invoke/unwind semantics are likely to change in future versions.</p>
3216
Chris Lattner00950542001-06-06 20:29:01 +00003217<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +00003218<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +00003219 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003220 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +00003221 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003222 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +00003223</pre>
Chris Lattner35eca582004-10-16 18:04:13 +00003224
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003225</div>
Chris Lattner35eca582004-10-16 18:04:13 +00003226
Chris Lattner27f71f22003-09-03 00:41:47 +00003227<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +00003228
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003229<h4>
3230 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3231</h4>
Chris Lattner35eca582004-10-16 18:04:13 +00003232
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003233<div>
Chris Lattner35eca582004-10-16 18:04:13 +00003234
Chris Lattner27f71f22003-09-03 00:41:47 +00003235<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003236<pre>
3237 unwind
3238</pre>
3239
Chris Lattner27f71f22003-09-03 00:41:47 +00003240<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003241<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003242 at the first callee in the dynamic call stack which used
3243 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3244 This is primarily used to implement exception handling.</p>
Chris Lattner35eca582004-10-16 18:04:13 +00003245
Chris Lattner27f71f22003-09-03 00:41:47 +00003246<h5>Semantics:</h5>
Chris Lattner72ed2002008-04-19 21:01:16 +00003247<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003248 immediately halt. The dynamic call stack is then searched for the
3249 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3250 Once found, execution continues at the "exceptional" destination block
3251 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3252 instruction in the dynamic call chain, undefined behavior results.</p>
3253
Chris Lattnerdf7a6802010-01-15 18:08:37 +00003254<p>Note that the code generator does not yet completely support unwind, and
3255that the invoke/unwind semantics are likely to change in future versions.</p>
3256
Misha Brukman9d0919f2003-11-08 01:05:38 +00003257</div>
Chris Lattner35eca582004-10-16 18:04:13 +00003258
3259<!-- _______________________________________________________________________ -->
3260
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003261<h4>
3262 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3263</h4>
Chris Lattner35eca582004-10-16 18:04:13 +00003264
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003265<div>
Chris Lattner35eca582004-10-16 18:04:13 +00003266
3267<h5>Syntax:</h5>
3268<pre>
3269 unreachable
3270</pre>
3271
3272<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003273<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003274 instruction is used to inform the optimizer that a particular portion of the
3275 code is not reachable. This can be used to indicate that the code after a
3276 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +00003277
3278<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +00003279<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003280
Chris Lattner35eca582004-10-16 18:04:13 +00003281</div>
3282
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003283</div>
3284
Chris Lattner00950542001-06-06 20:29:01 +00003285<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003286<h3>
3287 <a name="binaryops">Binary Operations</a>
3288</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003289
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003290<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003291
3292<p>Binary operators are used to do most of the computation in a program. They
3293 require two operands of the same type, execute an operation on them, and
3294 produce a single value. The operands might represent multiple data, as is
3295 the case with the <a href="#t_vector">vector</a> data type. The result value
3296 has the same type as its operands.</p>
3297
Misha Brukman9d0919f2003-11-08 01:05:38 +00003298<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003299
Chris Lattner00950542001-06-06 20:29:01 +00003300<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003301<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003302 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003303</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003304
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003305<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003306
Chris Lattner00950542001-06-06 20:29:01 +00003307<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003308<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003309 &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 +00003310 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3311 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3312 &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 +00003313</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003314
Chris Lattner00950542001-06-06 20:29:01 +00003315<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003316<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003317
Chris Lattner00950542001-06-06 20:29:01 +00003318<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003319<p>The two arguments to the '<tt>add</tt>' instruction must
3320 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3321 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003322
Chris Lattner00950542001-06-06 20:29:01 +00003323<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003324<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003325
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003326<p>If the sum has unsigned overflow, the result returned is the mathematical
3327 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003328
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003329<p>Because LLVM integers use a two's complement representation, this instruction
3330 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003331
Dan Gohman08d012e2009-07-22 22:44:56 +00003332<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3333 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3334 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +00003335 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3336 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003337
Chris Lattner00950542001-06-06 20:29:01 +00003338<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003339<pre>
3340 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00003341</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003342
Misha Brukman9d0919f2003-11-08 01:05:38 +00003343</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003344
Chris Lattner00950542001-06-06 20:29:01 +00003345<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003346<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003347 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003348</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003349
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003350<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003351
3352<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003353<pre>
3354 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3355</pre>
3356
3357<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003358<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3359
3360<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003361<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003362 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3363 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003364
3365<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003366<p>The value produced is the floating point sum of the two operands.</p>
3367
3368<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003369<pre>
3370 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3371</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003372
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003373</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003374
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003375<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003376<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003377 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003378</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003379
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003380<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003381
Chris Lattner00950542001-06-06 20:29:01 +00003382<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003383<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +00003384 &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 +00003385 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3386 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3387 &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 +00003388</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003389
Chris Lattner00950542001-06-06 20:29:01 +00003390<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003391<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003392 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003393
3394<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003395 '<tt>neg</tt>' instruction present in most other intermediate
3396 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003397
Chris Lattner00950542001-06-06 20:29:01 +00003398<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003399<p>The two arguments to the '<tt>sub</tt>' instruction must
3400 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3401 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003402
Chris Lattner00950542001-06-06 20:29:01 +00003403<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003404<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003405
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003406<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003407 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3408 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003409
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003410<p>Because LLVM integers use a two's complement representation, this instruction
3411 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003412
Dan Gohman08d012e2009-07-22 22:44:56 +00003413<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3414 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3415 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +00003416 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3417 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003418
Chris Lattner00950542001-06-06 20:29:01 +00003419<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +00003420<pre>
3421 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003422 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +00003423</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003424
Misha Brukman9d0919f2003-11-08 01:05:38 +00003425</div>
Chris Lattner5568e942008-05-20 20:48:21 +00003426
Chris Lattner00950542001-06-06 20:29:01 +00003427<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003428<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003429 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003430</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003431
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003432<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003433
3434<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003435<pre>
3436 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3437</pre>
3438
3439<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003440<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003441 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003442
3443<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003444 '<tt>fneg</tt>' instruction present in most other intermediate
3445 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003446
3447<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +00003448<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003449 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3450 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003451
3452<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003453<p>The value produced is the floating point difference of the two operands.</p>
3454
3455<h5>Example:</h5>
3456<pre>
3457 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3458 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3459</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003460
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003461</div>
3462
3463<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003464<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003465 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003466</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003467
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003468<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003469
Chris Lattner00950542001-06-06 20:29:01 +00003470<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003471<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +00003472 &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 +00003473 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3474 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3475 &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 +00003476</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003477
Chris Lattner00950542001-06-06 20:29:01 +00003478<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003479<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003480
Chris Lattner00950542001-06-06 20:29:01 +00003481<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003482<p>The two arguments to the '<tt>mul</tt>' instruction must
3483 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3484 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003485
Chris Lattner00950542001-06-06 20:29:01 +00003486<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003487<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003488
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003489<p>If the result of the multiplication has unsigned overflow, the result
3490 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3491 width of the result.</p>
3492
3493<p>Because LLVM integers use a two's complement representation, and the result
3494 is the same width as the operands, this instruction returns the correct
3495 result for both signed and unsigned integers. If a full product
3496 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3497 be sign-extended or zero-extended as appropriate to the width of the full
3498 product.</p>
3499
Dan Gohman08d012e2009-07-22 22:44:56 +00003500<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3501 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3502 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +00003503 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3504 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003505
Chris Lattner00950542001-06-06 20:29:01 +00003506<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003507<pre>
3508 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +00003509</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003510
Misha Brukman9d0919f2003-11-08 01:05:38 +00003511</div>
Chris Lattner5568e942008-05-20 20:48:21 +00003512
Chris Lattner00950542001-06-06 20:29:01 +00003513<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003514<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003515 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003516</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003517
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003518<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003519
3520<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003521<pre>
3522 &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 +00003523</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003524
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003525<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003526<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003527
3528<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003529<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003530 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3531 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003532
3533<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003534<p>The value produced is the floating point product of the two operands.</p>
3535
3536<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003537<pre>
3538 &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003539</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003540
Dan Gohmanae3a0be2009-06-04 22:49:04 +00003541</div>
3542
3543<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003544<h4>
3545 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3546</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003547
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003548<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003549
Reid Spencer1628cec2006-10-26 06:15:43 +00003550<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003551<pre>
Chris Lattner35bda892011-02-06 21:44:57 +00003552 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3553 &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 +00003554</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003555
Reid Spencer1628cec2006-10-26 06:15:43 +00003556<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003557<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003558
Reid Spencer1628cec2006-10-26 06:15:43 +00003559<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003560<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003561 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3562 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003563
Reid Spencer1628cec2006-10-26 06:15:43 +00003564<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +00003565<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003566
Chris Lattner5ec89832008-01-28 00:36:27 +00003567<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003568 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3569
Chris Lattner5ec89832008-01-28 00:36:27 +00003570<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003571
Chris Lattner35bda892011-02-06 21:44:57 +00003572<p>If the <tt>exact</tt> keyword is present, the result value of the
3573 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3574 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3575
3576
Reid Spencer1628cec2006-10-26 06:15:43 +00003577<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003578<pre>
3579 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00003580</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003581
Reid Spencer1628cec2006-10-26 06:15:43 +00003582</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003583
Reid Spencer1628cec2006-10-26 06:15:43 +00003584<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003585<h4>
3586 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3587</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003588
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003589<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003590
Reid Spencer1628cec2006-10-26 06:15:43 +00003591<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003592<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003593 &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 +00003594 &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 +00003595</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003596
Reid Spencer1628cec2006-10-26 06:15:43 +00003597<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003598<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003599
Reid Spencer1628cec2006-10-26 06:15:43 +00003600<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003601<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003602 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3603 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003604
Reid Spencer1628cec2006-10-26 06:15:43 +00003605<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003606<p>The value produced is the signed integer quotient of the two operands rounded
3607 towards zero.</p>
3608
Chris Lattner5ec89832008-01-28 00:36:27 +00003609<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003610 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3611
Chris Lattner5ec89832008-01-28 00:36:27 +00003612<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003613 undefined behavior; this is a rare case, but can occur, for example, by doing
3614 a 32-bit division of -2147483648 by -1.</p>
3615
Dan Gohman9c5beed2009-07-22 00:04:19 +00003616<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +00003617 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohman38da9272010-07-11 00:08:34 +00003618 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +00003619
Reid Spencer1628cec2006-10-26 06:15:43 +00003620<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003621<pre>
3622 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +00003623</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003624
Reid Spencer1628cec2006-10-26 06:15:43 +00003625</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003626
Reid Spencer1628cec2006-10-26 06:15:43 +00003627<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003628<h4>
3629 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3630</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003631
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003632<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003633
Chris Lattner00950542001-06-06 20:29:01 +00003634<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003635<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +00003636 &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 +00003637</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003638
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003639<h5>Overview:</h5>
3640<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003641
Chris Lattner261efe92003-11-25 01:02:51 +00003642<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00003643<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003644 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3645 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003646
Chris Lattner261efe92003-11-25 01:02:51 +00003647<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +00003648<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003649
Chris Lattner261efe92003-11-25 01:02:51 +00003650<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003651<pre>
3652 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00003653</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003654
Chris Lattner261efe92003-11-25 01:02:51 +00003655</div>
Chris Lattner5568e942008-05-20 20:48:21 +00003656
Chris Lattner261efe92003-11-25 01:02:51 +00003657<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003658<h4>
3659 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3660</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003661
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003662<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003663
Reid Spencer0a783f72006-11-02 01:53:59 +00003664<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003665<pre>
3666 &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 +00003667</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003668
Reid Spencer0a783f72006-11-02 01:53:59 +00003669<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003670<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3671 division of its two arguments.</p>
3672
Reid Spencer0a783f72006-11-02 01:53:59 +00003673<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003674<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003675 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3676 values. Both arguments must have identical types.</p>
3677
Reid Spencer0a783f72006-11-02 01:53:59 +00003678<h5>Semantics:</h5>
3679<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003680 This instruction always performs an unsigned division to get the
3681 remainder.</p>
3682
Chris Lattner5ec89832008-01-28 00:36:27 +00003683<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003684 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3685
Chris Lattner5ec89832008-01-28 00:36:27 +00003686<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003687
Reid Spencer0a783f72006-11-02 01:53:59 +00003688<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003689<pre>
3690 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00003691</pre>
3692
3693</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003694
Reid Spencer0a783f72006-11-02 01:53:59 +00003695<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003696<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003697 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003698</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003699
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003700<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003701
Chris Lattner261efe92003-11-25 01:02:51 +00003702<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003703<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +00003704 &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 +00003705</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003706
Chris Lattner261efe92003-11-25 01:02:51 +00003707<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003708<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3709 division of its two operands. This instruction can also take
3710 <a href="#t_vector">vector</a> versions of the values in which case the
3711 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +00003712
Chris Lattner261efe92003-11-25 01:02:51 +00003713<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003714<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003715 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3716 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003717
Chris Lattner261efe92003-11-25 01:02:51 +00003718<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +00003719<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +00003720 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3721 <i>modulo</i> operator (where the result is either zero or has the same sign
3722 as the divisor, <tt>op2</tt>) of a value.
3723 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003724 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3725 Math Forum</a>. For a table of how this is implemented in various languages,
3726 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3727 Wikipedia: modulo operation</a>.</p>
3728
Chris Lattner5ec89832008-01-28 00:36:27 +00003729<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003730 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3731
Chris Lattner5ec89832008-01-28 00:36:27 +00003732<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003733 Overflow also leads to undefined behavior; this is a rare case, but can
3734 occur, for example, by taking the remainder of a 32-bit division of
3735 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3736 lets srem be implemented using instructions that return both the result of
3737 the division and the remainder.)</p>
3738
Chris Lattner261efe92003-11-25 01:02:51 +00003739<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003740<pre>
3741 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +00003742</pre>
3743
3744</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003745
Reid Spencer0a783f72006-11-02 01:53:59 +00003746<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003747<h4>
3748 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3749</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003750
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003751<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003752
Reid Spencer0a783f72006-11-02 01:53:59 +00003753<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003754<pre>
3755 &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 +00003756</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003757
Reid Spencer0a783f72006-11-02 01:53:59 +00003758<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003759<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3760 its two operands.</p>
3761
Reid Spencer0a783f72006-11-02 01:53:59 +00003762<h5>Arguments:</h5>
3763<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003764 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3765 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003766
Reid Spencer0a783f72006-11-02 01:53:59 +00003767<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003768<p>This instruction returns the <i>remainder</i> of a division. The remainder
3769 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003770
Reid Spencer0a783f72006-11-02 01:53:59 +00003771<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003772<pre>
3773 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +00003774</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003775
Misha Brukman9d0919f2003-11-08 01:05:38 +00003776</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +00003777
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003778</div>
3779
Reid Spencer8e11bf82007-02-02 13:57:07 +00003780<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003781<h3>
3782 <a name="bitwiseops">Bitwise Binary Operations</a>
3783</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003784
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003785<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003786
3787<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3788 program. They are generally very efficient instructions and can commonly be
3789 strength reduced from other instructions. They require two operands of the
3790 same type, execute an operation on them, and produce a single value. The
3791 resulting value is the same type as its operands.</p>
3792
Reid Spencer569f2fa2007-01-31 21:39:12 +00003793<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003794<h4>
3795 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3796</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003797
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003798<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003799
Reid Spencer569f2fa2007-01-31 21:39:12 +00003800<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003801<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +00003802 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3803 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3804 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3805 &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 +00003806</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003807
Reid Spencer569f2fa2007-01-31 21:39:12 +00003808<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003809<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3810 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003811
Reid Spencer569f2fa2007-01-31 21:39:12 +00003812<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003813<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3814 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3815 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003816
Reid Spencer569f2fa2007-01-31 21:39:12 +00003817<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003818<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3819 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3820 is (statically or dynamically) negative or equal to or larger than the number
3821 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3822 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3823 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003824
Chris Lattnerf067d582011-02-07 16:40:21 +00003825<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3826 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
Chris Lattner66298c12011-02-09 16:44:44 +00003827 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnerf067d582011-02-07 16:40:21 +00003828 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3829 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3830 they would if the shift were expressed as a mul instruction with the same
3831 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3832
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003833<h5>Example:</h5>
3834<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003835 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3836 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3837 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003838 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +00003839 &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 +00003840</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003841
Reid Spencer569f2fa2007-01-31 21:39:12 +00003842</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003843
Reid Spencer569f2fa2007-01-31 21:39:12 +00003844<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003845<h4>
3846 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
3847</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003848
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003849<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003850
Reid Spencer569f2fa2007-01-31 21:39:12 +00003851<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003852<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +00003853 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3854 &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 +00003855</pre>
3856
3857<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003858<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3859 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003860
3861<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003862<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003863 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3864 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003865
3866<h5>Semantics:</h5>
3867<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003868 significant bits of the result will be filled with zero bits after the shift.
3869 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3870 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3871 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3872 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003873
Chris Lattnerf067d582011-02-07 16:40:21 +00003874<p>If the <tt>exact</tt> keyword is present, the result value of the
3875 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3876 shifted out are non-zero.</p>
3877
3878
Reid Spencer569f2fa2007-01-31 21:39:12 +00003879<h5>Example:</h5>
3880<pre>
3881 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3882 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3883 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3884 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003885 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +00003886 &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 +00003887</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003888
Reid Spencer569f2fa2007-01-31 21:39:12 +00003889</div>
3890
Reid Spencer8e11bf82007-02-02 13:57:07 +00003891<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003892<h4>
3893 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
3894</h4>
3895
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003896<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003897
3898<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003899<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +00003900 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3901 &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 +00003902</pre>
3903
3904<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003905<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3906 operand shifted to the right a specified number of bits with sign
3907 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003908
3909<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003910<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003911 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3912 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003913
3914<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003915<p>This instruction always performs an arithmetic shift right operation, The
3916 most significant bits of the result will be filled with the sign bit
3917 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3918 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3919 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3920 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +00003921
Chris Lattnerf067d582011-02-07 16:40:21 +00003922<p>If the <tt>exact</tt> keyword is present, the result value of the
3923 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3924 shifted out are non-zero.</p>
3925
Reid Spencer569f2fa2007-01-31 21:39:12 +00003926<h5>Example:</h5>
3927<pre>
3928 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3929 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3930 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3931 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +00003932 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +00003933 &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 +00003934</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003935
Reid Spencer569f2fa2007-01-31 21:39:12 +00003936</div>
3937
Chris Lattner00950542001-06-06 20:29:01 +00003938<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00003939<h4>
3940 <a name="i_and">'<tt>and</tt>' Instruction</a>
3941</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00003942
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00003943<div>
Chris Lattner5568e942008-05-20 20:48:21 +00003944
Chris Lattner00950542001-06-06 20:29:01 +00003945<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003946<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +00003947 &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 +00003948</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00003949
Chris Lattner00950542001-06-06 20:29:01 +00003950<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003951<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3952 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003953
Chris Lattner00950542001-06-06 20:29:01 +00003954<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00003955<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003956 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3957 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00003958
Chris Lattner00950542001-06-06 20:29:01 +00003959<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003960<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003961
Misha Brukman9d0919f2003-11-08 01:05:38 +00003962<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +00003963 <tbody>
3964 <tr>
3965 <td>In0</td>
3966 <td>In1</td>
3967 <td>Out</td>
3968 </tr>
3969 <tr>
3970 <td>0</td>
3971 <td>0</td>
3972 <td>0</td>
3973 </tr>
3974 <tr>
3975 <td>0</td>
3976 <td>1</td>
3977 <td>0</td>
3978 </tr>
3979 <tr>
3980 <td>1</td>
3981 <td>0</td>
3982 <td>0</td>
3983 </tr>
3984 <tr>
3985 <td>1</td>
3986 <td>1</td>
3987 <td>1</td>
3988 </tr>
3989 </tbody>
3990</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00003991
Chris Lattner00950542001-06-06 20:29:01 +00003992<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00003993<pre>
3994 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00003995 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3996 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +00003997</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +00003998</div>
Chris Lattner00950542001-06-06 20:29:01 +00003999<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004000<h4>
4001 <a name="i_or">'<tt>or</tt>' Instruction</a>
4002</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00004003
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004004<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004005
4006<h5>Syntax:</h5>
4007<pre>
4008 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4009</pre>
4010
4011<h5>Overview:</h5>
4012<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4013 two operands.</p>
4014
4015<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004016<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004017 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4018 values. Both arguments must have identical types.</p>
4019
Chris Lattner00950542001-06-06 20:29:01 +00004020<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004021<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004022
Chris Lattner261efe92003-11-25 01:02:51 +00004023<table border="1" cellspacing="0" cellpadding="4">
4024 <tbody>
4025 <tr>
4026 <td>In0</td>
4027 <td>In1</td>
4028 <td>Out</td>
4029 </tr>
4030 <tr>
4031 <td>0</td>
4032 <td>0</td>
4033 <td>0</td>
4034 </tr>
4035 <tr>
4036 <td>0</td>
4037 <td>1</td>
4038 <td>1</td>
4039 </tr>
4040 <tr>
4041 <td>1</td>
4042 <td>0</td>
4043 <td>1</td>
4044 </tr>
4045 <tr>
4046 <td>1</td>
4047 <td>1</td>
4048 <td>1</td>
4049 </tr>
4050 </tbody>
4051</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004052
Chris Lattner00950542001-06-06 20:29:01 +00004053<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004054<pre>
4055 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004056 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4057 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +00004058</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004059
Misha Brukman9d0919f2003-11-08 01:05:38 +00004060</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004061
Chris Lattner00950542001-06-06 20:29:01 +00004062<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004063<h4>
4064 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4065</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004066
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004067<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004068
Chris Lattner00950542001-06-06 20:29:01 +00004069<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004070<pre>
4071 &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 +00004072</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004073
Chris Lattner00950542001-06-06 20:29:01 +00004074<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004075<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4076 its two operands. The <tt>xor</tt> is used to implement the "one's
4077 complement" operation, which is the "~" operator in C.</p>
4078
Chris Lattner00950542001-06-06 20:29:01 +00004079<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004080<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004081 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4082 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00004083
Chris Lattner00950542001-06-06 20:29:01 +00004084<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004085<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004086
Chris Lattner261efe92003-11-25 01:02:51 +00004087<table border="1" cellspacing="0" cellpadding="4">
4088 <tbody>
4089 <tr>
4090 <td>In0</td>
4091 <td>In1</td>
4092 <td>Out</td>
4093 </tr>
4094 <tr>
4095 <td>0</td>
4096 <td>0</td>
4097 <td>0</td>
4098 </tr>
4099 <tr>
4100 <td>0</td>
4101 <td>1</td>
4102 <td>1</td>
4103 </tr>
4104 <tr>
4105 <td>1</td>
4106 <td>0</td>
4107 <td>1</td>
4108 </tr>
4109 <tr>
4110 <td>1</td>
4111 <td>1</td>
4112 <td>0</td>
4113 </tr>
4114 </tbody>
4115</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004116
Chris Lattner00950542001-06-06 20:29:01 +00004117<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004118<pre>
4119 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004120 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4121 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4122 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +00004123</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004124
Misha Brukman9d0919f2003-11-08 01:05:38 +00004125</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004126
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004127</div>
4128
Chris Lattner00950542001-06-06 20:29:01 +00004129<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004130<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +00004131 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004132</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +00004133
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004134<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004135
4136<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004137 target-independent manner. These instructions cover the element-access and
4138 vector-specific operations needed to process vectors effectively. While LLVM
4139 does directly support these vector operations, many sophisticated algorithms
4140 will want to use target-specific intrinsics to take full advantage of a
4141 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004142
Chris Lattner3df241e2006-04-08 23:07:04 +00004143<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004144<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004145 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004146</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004147
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004148<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004149
4150<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004151<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004152 &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 +00004153</pre>
4154
4155<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004156<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4157 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004158
4159
4160<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004161<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4162 of <a href="#t_vector">vector</a> type. The second operand is an index
4163 indicating the position from which to extract the element. The index may be
4164 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004165
4166<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004167<p>The result is a scalar of the same type as the element type of
4168 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4169 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4170 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004171
4172<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004173<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +00004174 &lt;result&gt; = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattner3df241e2006-04-08 23:07:04 +00004175</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +00004176
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004177</div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004178
4179<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004180<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004181 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004182</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004183
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004184<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004185
4186<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004187<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +00004188 &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 +00004189</pre>
4190
4191<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004192<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4193 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004194
4195<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004196<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4197 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4198 whose type must equal the element type of the first operand. The third
4199 operand is an index indicating the position at which to insert the value.
4200 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004201
4202<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004203<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4204 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4205 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4206 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004207
4208<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004209<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +00004210 &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 +00004211</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004212
Chris Lattner3df241e2006-04-08 23:07:04 +00004213</div>
4214
4215<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004216<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004217 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004218</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +00004219
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004220<div>
Chris Lattner3df241e2006-04-08 23:07:04 +00004221
4222<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004223<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +00004224 &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 +00004225</pre>
4226
4227<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004228<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4229 from two input vectors, returning a vector with the same element type as the
4230 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004231
4232<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004233<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4234 with types that match each other. The third argument is a shuffle mask whose
4235 element type is always 'i32'. The result of the instruction is a vector
4236 whose length is the same as the shuffle mask and whose element type is the
4237 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004238
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004239<p>The shuffle mask operand is required to be a constant vector with either
4240 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004241
4242<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004243<p>The elements of the two input vectors are numbered from left to right across
4244 both of the vectors. The shuffle mask operand specifies, for each element of
4245 the result vector, which element of the two input vectors the result element
4246 gets. The element selector may be undef (meaning "don't care") and the
4247 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +00004248
4249<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +00004250<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004251 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen6f1cc772007-04-22 01:17:39 +00004252 &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 +00004253 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
Reid Spencerca86e162006-12-31 07:07:53 +00004254 &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 +00004255 &lt;result&gt; = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
Mon P Wangaeb06d22008-11-10 04:46:22 +00004256 &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 +00004257 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Mon P Wangaeb06d22008-11-10 04:46:22 +00004258 &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 +00004259</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +00004260
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004261</div>
Tanya Lattner09474292006-04-14 19:24:33 +00004262
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004263</div>
4264
Chris Lattner3df241e2006-04-08 23:07:04 +00004265<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004266<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004267 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004268</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004269
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004270<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004271
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004272<p>LLVM supports several instructions for working with
4273 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004274
Dan Gohmana334d5f2008-05-12 23:51:09 +00004275<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004276<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004277 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004278</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004279
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004280<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004281
4282<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004283<pre>
4284 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4285</pre>
4286
4287<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004288<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4289 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004290
4291<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004292<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +00004293 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004294 <a href="#t_array">array</a> type. The operands are constant indices to
4295 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004296 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +00004297 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4298 <ul>
4299 <li>Since the value being indexed is not a pointer, the first index is
4300 omitted and assumed to be zero.</li>
4301 <li>At least one index must be specified.</li>
4302 <li>Not only struct indices but also array indices must be in
4303 bounds.</li>
4304 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004305
4306<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004307<p>The result is the value at the position in the aggregate specified by the
4308 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004309
4310<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004311<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +00004312 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004313</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004314
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004315</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004316
4317<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004318<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004319 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004320</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004321
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004322<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004323
4324<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004325<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +00004326 &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 +00004327</pre>
4328
4329<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004330<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4331 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004332
4333<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004334<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +00004335 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004336 <a href="#t_array">array</a> type. The second operand is a first-class
4337 value to insert. The following operands are constant indices indicating
4338 the position at which to insert the value in a similar manner as indices in a
Frits van Bommel13242892010-12-05 20:54:38 +00004339 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004340 value to insert must have the same type as the value identified by the
4341 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004342
4343<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004344<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4345 that of <tt>val</tt> except that the value at the position specified by the
4346 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004347
4348<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004349<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +00004350 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4351 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4352 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 <i>; yields {i32 1, float %val}</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004353</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004354
Dan Gohmana334d5f2008-05-12 23:51:09 +00004355</div>
4356
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004357</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +00004358
4359<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004360<h3>
Chris Lattner884a9702006-08-15 00:45:58 +00004361 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004362</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004363
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004364<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004365
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004366<p>A key design point of an SSA-based representation is how it represents
4367 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +00004368 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004369 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004370
Chris Lattner00950542001-06-06 20:29:01 +00004371<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004372<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004373 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004374</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004375
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004376<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004377
Chris Lattner00950542001-06-06 20:29:01 +00004378<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004379<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +00004380 &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 +00004381</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004382
Chris Lattner00950542001-06-06 20:29:01 +00004383<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00004384<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004385 currently executing function, to be automatically released when this function
4386 returns to its caller. The object is always allocated in the generic address
4387 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004388
Chris Lattner00950542001-06-06 20:29:01 +00004389<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004390<p>The '<tt>alloca</tt>' instruction
4391 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4392 runtime stack, returning a pointer of the appropriate type to the program.
4393 If "NumElements" is specified, it is the number of elements allocated,
4394 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4395 specified, the value result of the allocation is guaranteed to be aligned to
4396 at least that boundary. If not specified, or if zero, the target can choose
4397 to align the allocation on any convenient boundary compatible with the
4398 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004399
Misha Brukman9d0919f2003-11-08 01:05:38 +00004400<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004401
Chris Lattner00950542001-06-06 20:29:01 +00004402<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +00004403<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004404 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4405 memory is automatically released when the function returns. The
4406 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4407 variables that must have an address available. When the function returns
4408 (either with the <tt><a href="#i_ret">ret</a></tt>
4409 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4410 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004411
Chris Lattner00950542001-06-06 20:29:01 +00004412<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004413<pre>
Dan Gohman81e21672009-01-04 23:49:44 +00004414 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4415 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4416 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4417 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +00004418</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004419
Misha Brukman9d0919f2003-11-08 01:05:38 +00004420</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +00004421
Chris Lattner00950542001-06-06 20:29:01 +00004422<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004423<h4>
4424 <a name="i_load">'<tt>load</tt>' Instruction</a>
4425</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004426
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004427<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004428
Chris Lattner2b7d3202002-05-06 03:03:22 +00004429<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004430<pre>
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004431 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4432 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4433 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004434</pre>
4435
Chris Lattner2b7d3202002-05-06 03:03:22 +00004436<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004437<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004438
Chris Lattner2b7d3202002-05-06 03:03:22 +00004439<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004440<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4441 from which to load. The pointer must point to
4442 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4443 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00004444 number or order of execution of this <tt>load</tt> with other <a
4445 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004446
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004447<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004448 operation (that is, the alignment of the memory address). A value of 0 or an
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004449 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004450 alignment for the target. It is the responsibility of the code emitter to
4451 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004452 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004453 produce less efficient code. An alignment of 1 is always safe.</p>
4454
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004455<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4456 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004457 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +00004458 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4459 and code generator that this load is not expected to be reused in the cache.
4460 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004461 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +00004462
Chris Lattner2b7d3202002-05-06 03:03:22 +00004463<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004464<p>The location of memory pointed to is loaded. If the value being loaded is of
4465 scalar type then the number of bytes read does not exceed the minimum number
4466 of bytes needed to hold all bits of the type. For example, loading an
4467 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4468 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4469 is undefined if the value was not originally written using a store of the
4470 same type.</p>
4471
Chris Lattner2b7d3202002-05-06 03:03:22 +00004472<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004473<pre>
4474 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4475 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004476 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00004477</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004478
Misha Brukman9d0919f2003-11-08 01:05:38 +00004479</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004480
Chris Lattner2b7d3202002-05-06 03:03:22 +00004481<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004482<h4>
4483 <a name="i_store">'<tt>store</tt>' Instruction</a>
4484</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004485
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004486<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004487
Chris Lattner2b7d3202002-05-06 03:03:22 +00004488<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004489<pre>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004490 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>
4491 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 +00004492</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004493
Chris Lattner2b7d3202002-05-06 03:03:22 +00004494<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004495<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004496
Chris Lattner2b7d3202002-05-06 03:03:22 +00004497<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004498<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4499 and an address at which to store it. The type of the
4500 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4501 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00004502 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4503 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4504 order of execution of this <tt>store</tt> with other <a
4505 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004506
4507<p>The optional constant "align" argument specifies the alignment of the
4508 operation (that is, the alignment of the memory address). A value of 0 or an
4509 omitted "align" argument means that the operation has the preferential
4510 alignment for the target. It is the responsibility of the code emitter to
4511 ensure that the alignment information is correct. Overestimating the
4512 alignment results in an undefined behavior. Underestimating the alignment may
4513 produce less efficient code. An alignment of 1 is always safe.</p>
4514
David Greene8939b0d2010-02-16 20:50:18 +00004515<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004516 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004517 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +00004518 instruction tells the optimizer and code generator that this load is
4519 not expected to be reused in the cache. The code generator may
4520 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00004521 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +00004522
4523
Chris Lattner261efe92003-11-25 01:02:51 +00004524<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004525<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4526 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4527 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4528 does not exceed the minimum number of bytes needed to hold all bits of the
4529 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4530 writing a value of a type like <tt>i20</tt> with a size that is not an
4531 integral number of bytes, it is unspecified what happens to the extra bits
4532 that do not belong to the type, but they will typically be overwritten.</p>
4533
Chris Lattner2b7d3202002-05-06 03:03:22 +00004534<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004535<pre>
4536 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +00004537 store i32 3, i32* %ptr <i>; yields {void}</i>
4538 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +00004539</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004540
Reid Spencer47ce1792006-11-09 21:15:49 +00004541</div>
4542
Chris Lattner2b7d3202002-05-06 03:03:22 +00004543<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004544<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004545 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004546</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004547
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004548<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004549
Chris Lattner7faa8832002-04-14 06:13:44 +00004550<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004551<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004552 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohmandd8004d2009-07-27 21:53:46 +00004553 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004554</pre>
4555
Chris Lattner7faa8832002-04-14 06:13:44 +00004556<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004557<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004558 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4559 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004560
Chris Lattner7faa8832002-04-14 06:13:44 +00004561<h5>Arguments:</h5>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004562<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +00004563 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004564 elements of the aggregate object are indexed. The interpretation of each
4565 index is dependent on the type being indexed into. The first index always
4566 indexes the pointer value given as the first argument, the second index
4567 indexes a value of the type pointed to (not necessarily the value directly
4568 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004569 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +00004570 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004571 can never be pointers, since that would require loading the pointer before
4572 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004573
4574<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner61c70e92010-08-28 04:09:24 +00004575 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +00004576 integer <b>constants</b> are allowed. When indexing into an array, pointer
4577 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnerc8eef442009-07-29 06:44:13 +00004578 constant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004579
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004580<p>For example, let's consider a C code fragment and how it gets compiled to
4581 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004582
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004583<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004584struct RT {
4585 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +00004586 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004587 char C;
4588};
4589struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +00004590 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004591 double Y;
4592 struct RT Z;
4593};
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004594
Chris Lattnercabc8462007-05-29 15:43:56 +00004595int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004596 return &amp;s[1].Z.B[5][13];
4597}
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004598</pre>
4599
Misha Brukman9d0919f2003-11-08 01:05:38 +00004600<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004601
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00004602<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +00004603%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4604%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004605
Dan Gohman4df605b2009-07-25 02:23:48 +00004606define i32* @foo(%ST* %s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +00004607entry:
4608 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4609 ret i32* %reg
4610}
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004611</pre>
4612
Chris Lattner7faa8832002-04-14 06:13:44 +00004613<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00004614<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004615 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4616 }</tt>' type, a structure. The second index indexes into the third element
4617 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4618 i8 }</tt>' type, another structure. The third index indexes into the second
4619 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4620 array. The two dimensions of the array are subscripted into, yielding an
4621 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4622 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004623
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004624<p>Note that it is perfectly legal to index partially through a structure,
4625 returning a pointer to an inner element. Because of this, the LLVM code for
4626 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004627
4628<pre>
Dan Gohman4df605b2009-07-25 02:23:48 +00004629 define i32* @foo(%ST* %s) {
Reid Spencerca86e162006-12-31 07:07:53 +00004630 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +00004631 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4632 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerca86e162006-12-31 07:07:53 +00004633 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4634 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4635 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004636 }
Chris Lattner6536cfe2002-05-06 22:08:29 +00004637</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +00004638
Dan Gohmandd8004d2009-07-27 21:53:46 +00004639<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +00004640 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4641 base pointer is not an <i>in bounds</i> address of an allocated object,
4642 or if any of the addresses that would be formed by successive addition of
4643 the offsets implied by the indices to the base address with infinitely
4644 precise arithmetic are not an <i>in bounds</i> address of that allocated
4645 object. The <i>in bounds</i> addresses for an allocated object are all
4646 the addresses that point into the object, plus the address one byte past
4647 the end.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +00004648
4649<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4650 the base address with silently-wrapping two's complement arithmetic, and
4651 the result value of the <tt>getelementptr</tt> may be outside the object
4652 pointed to by the base pointer. The result value may not necessarily be
4653 used to access memory though, even if it happens to point into allocated
4654 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4655 section for more information.</p>
4656
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004657<p>The getelementptr instruction is often confusing. For some more insight into
4658 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +00004659
Chris Lattner7faa8832002-04-14 06:13:44 +00004660<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004661<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004662 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004663 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4664 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +00004665 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +00004666 <i>; yields i8*:eptr</i>
4667 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +00004668 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +00004669 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004670</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004671
Chris Lattnerf74d5c72004-04-05 01:30:49 +00004672</div>
Reid Spencer47ce1792006-11-09 21:15:49 +00004673
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004674</div>
4675
Chris Lattner00950542001-06-06 20:29:01 +00004676<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004677<h3>
4678 <a name="convertops">Conversion Operations</a>
4679</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004680
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004681<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004682
Reid Spencer2fd21e62006-11-08 01:18:52 +00004683<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004684 which all take a single operand and a type. They perform various bit
4685 conversions on the operand.</p>
4686
Chris Lattner6536cfe2002-05-06 22:08:29 +00004687<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004688<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004689 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004690</h4>
4691
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004692<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004693
4694<h5>Syntax:</h5>
4695<pre>
4696 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4697</pre>
4698
4699<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004700<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4701 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004702
4703<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004704<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
4705 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4706 of the same number of integers.
4707 The bit size of the <tt>value</tt> must be larger than
4708 the bit size of the destination type, <tt>ty2</tt>.
4709 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004710
4711<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004712<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4713 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4714 source size must be larger than the destination size, <tt>trunc</tt> cannot
4715 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004716
4717<h5>Example:</h5>
4718<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004719 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4720 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
4721 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
4722 %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 +00004723</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004724
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004725</div>
4726
4727<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004728<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004729 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004730</h4>
4731
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004732<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004733
4734<h5>Syntax:</h5>
4735<pre>
4736 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4737</pre>
4738
4739<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004740<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004741 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004742
4743
4744<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +00004745<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
4746 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4747 of the same number of integers.
4748 The bit size of the <tt>value</tt> must be smaller than
4749 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004750 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004751
4752<h5>Semantics:</h5>
4753<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004754 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004755
Reid Spencerb5929522007-01-12 15:46:11 +00004756<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004757
4758<h5>Example:</h5>
4759<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004760 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00004761 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +00004762 %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 +00004763</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004764
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004765</div>
4766
4767<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004768<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004769 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004770</h4>
4771
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004772<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004773
4774<h5>Syntax:</h5>
4775<pre>
4776 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4777</pre>
4778
4779<h5>Overview:</h5>
4780<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4781
4782<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004783<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
4784 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4785 of the same number of integers.
4786 The bit size of the <tt>value</tt> must be smaller than
4787 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004788 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004789
4790<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004791<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4792 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4793 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004794
Reid Spencerc78f3372007-01-12 03:35:51 +00004795<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004796
4797<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004798<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004799 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +00004800 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +00004801 %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 +00004802</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004803
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004804</div>
4805
4806<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004807<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004808 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004809</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004810
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004811<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004812
4813<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004814<pre>
4815 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4816</pre>
4817
4818<h5>Overview:</h5>
4819<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004820 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004821
4822<h5>Arguments:</h5>
4823<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004824 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4825 to cast it to. The size of <tt>value</tt> must be larger than the size of
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004826 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004827 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004828
4829<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004830<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004831 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004832 <a href="#t_floating">floating point</a> type. If the value cannot fit
4833 within the destination type, <tt>ty2</tt>, then the results are
4834 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +00004835
4836<h5>Example:</h5>
4837<pre>
4838 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4839 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4840</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004841
Reid Spencer3fa91b02006-11-09 21:48:10 +00004842</div>
4843
4844<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004845<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004846 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004847</h4>
4848
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004849<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004850
4851<h5>Syntax:</h5>
4852<pre>
4853 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4854</pre>
4855
4856<h5>Overview:</h5>
4857<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004858 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004859
4860<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004861<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004862 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4863 a <a href="#t_floating">floating point</a> type to cast it to. The source
4864 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004865
4866<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00004867<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004868 <a href="#t_floating">floating point</a> type to a larger
4869 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4870 used to make a <i>no-op cast</i> because it always changes bits. Use
4871 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004872
4873<h5>Example:</h5>
4874<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +00004875 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
4876 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004877</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004878
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004879</div>
4880
4881<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004882<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +00004883 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004884</h4>
4885
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004886<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004887
4888<h5>Syntax:</h5>
4889<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +00004890 &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 +00004891</pre>
4892
4893<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +00004894<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004895 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004896
4897<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004898<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4899 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4900 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4901 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4902 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004903
4904<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004905<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004906 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4907 towards zero) unsigned integer value. If the value cannot fit
4908 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004909
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004910<h5>Example:</h5>
4911<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +00004912 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +00004913 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +00004914 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004915</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004916
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004917</div>
4918
4919<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004920<h4>
Reid Spencerd4448792006-11-09 23:03:26 +00004921 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004922</h4>
4923
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004924<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004925
4926<h5>Syntax:</h5>
4927<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00004928 &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 +00004929</pre>
4930
4931<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004932<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004933 <a href="#t_floating">floating point</a> <tt>value</tt> to
4934 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004935
Chris Lattner6536cfe2002-05-06 22:08:29 +00004936<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004937<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4938 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4939 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4940 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4941 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004942
Chris Lattner6536cfe2002-05-06 22:08:29 +00004943<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00004944<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004945 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4946 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4947 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004948
Chris Lattner33ba0d92001-07-09 00:26:23 +00004949<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00004950<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +00004951 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +00004952 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +00004953 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004954</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004955
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004956</div>
4957
4958<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004959<h4>
Reid Spencerd4448792006-11-09 23:03:26 +00004960 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004961</h4>
4962
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00004963<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004964
4965<h5>Syntax:</h5>
4966<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00004967 &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 +00004968</pre>
4969
4970<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00004971<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004972 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004973
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004974<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +00004975<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004976 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4977 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4978 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4979 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004980
4981<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +00004982<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004983 integer quantity and converts it to the corresponding floating point
4984 value. If the value cannot fit in the floating point value, the results are
4985 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004986
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004987<h5>Example:</h5>
4988<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00004989 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +00004990 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004991</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00004992
Reid Spencer9dee3ac2006-11-08 01:11:31 +00004993</div>
4994
4995<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004996<h4>
Reid Spencerd4448792006-11-09 23:03:26 +00004997 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00004998</h4>
4999
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005000<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005001
5002<h5>Syntax:</h5>
5003<pre>
Reid Spencerd4448792006-11-09 23:03:26 +00005004 &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 +00005005</pre>
5006
5007<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005008<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5009 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005010
5011<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +00005012<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005013 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5014 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5015 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5016 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005017
5018<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005019<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5020 quantity and converts it to the corresponding floating point value. If the
5021 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005022
5023<h5>Example:</h5>
5024<pre>
Reid Spencerca86e162006-12-31 07:07:53 +00005025 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +00005026 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005027</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005028
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005029</div>
5030
5031<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005032<h4>
Reid Spencer72679252006-11-11 21:00:47 +00005033 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005034</h4>
5035
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005036<div>
Reid Spencer72679252006-11-11 21:00:47 +00005037
5038<h5>Syntax:</h5>
5039<pre>
5040 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5041</pre>
5042
5043<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005044<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5045 the integer type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005046
5047<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005048<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5049 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5050 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005051
5052<h5>Semantics:</h5>
5053<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005054 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5055 truncating or zero extending that value to the size of the integer type. If
5056 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5057 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5058 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5059 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005060
5061<h5>Example:</h5>
5062<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005063 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5064 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +00005065</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005066
Reid Spencer72679252006-11-11 21:00:47 +00005067</div>
5068
5069<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005070<h4>
Reid Spencer72679252006-11-11 21:00:47 +00005071 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005072</h4>
5073
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005074<div>
Reid Spencer72679252006-11-11 21:00:47 +00005075
5076<h5>Syntax:</h5>
5077<pre>
5078 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5079</pre>
5080
5081<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005082<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5083 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005084
5085<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +00005086<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005087 value to cast, and a type to cast it to, which must be a
5088 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +00005089
5090<h5>Semantics:</h5>
5091<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005092 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5093 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5094 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5095 than the size of a pointer then a zero extension is done. If they are the
5096 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +00005097
5098<h5>Example:</h5>
5099<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005100 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
Gabor Greif6a292012009-10-28 09:21:30 +00005101 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5102 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +00005103</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005104
Reid Spencer72679252006-11-11 21:00:47 +00005105</div>
5106
5107<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005108<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005109 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005110</h4>
5111
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005112<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005113
5114<h5>Syntax:</h5>
5115<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005116 &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 +00005117</pre>
5118
5119<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005120<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005121 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005122
5123<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005124<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5125 non-aggregate first class value, and a type to cast it to, which must also be
5126 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5127 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5128 identical. If the source type is a pointer, the destination type must also be
5129 a pointer. This instruction supports bitwise conversion of vectors to
5130 integers and to vectors of other types (as long as they have the same
5131 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005132
5133<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +00005134<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005135 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5136 this conversion. The conversion is done as if the <tt>value</tt> had been
5137 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5138 be converted to other pointer types with this instruction. To convert
5139 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5140 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +00005141
5142<h5>Example:</h5>
5143<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005144 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +00005145 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005146 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +00005147</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005148
Misha Brukman9d0919f2003-11-08 01:05:38 +00005149</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005150
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005151</div>
5152
Reid Spencer2fd21e62006-11-08 01:18:52 +00005153<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005154<h3>
5155 <a name="otherops">Other Operations</a>
5156</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005157
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005158<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005159
5160<p>The instructions in this category are the "miscellaneous" instructions, which
5161 defy better classification.</p>
5162
Reid Spencerf3a70a62006-11-18 21:50:54 +00005163<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005164<h4>
5165 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5166</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005167
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005168<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005169
Reid Spencerf3a70a62006-11-18 21:50:54 +00005170<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005171<pre>
5172 &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 +00005173</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005174
Reid Spencerf3a70a62006-11-18 21:50:54 +00005175<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005176<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5177 boolean values based on comparison of its two integer, integer vector, or
5178 pointer operands.</p>
5179
Reid Spencerf3a70a62006-11-18 21:50:54 +00005180<h5>Arguments:</h5>
5181<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005182 the condition code indicating the kind of comparison to perform. It is not a
5183 value, just a keyword. The possible condition code are:</p>
5184
Reid Spencerf3a70a62006-11-18 21:50:54 +00005185<ol>
5186 <li><tt>eq</tt>: equal</li>
5187 <li><tt>ne</tt>: not equal </li>
5188 <li><tt>ugt</tt>: unsigned greater than</li>
5189 <li><tt>uge</tt>: unsigned greater or equal</li>
5190 <li><tt>ult</tt>: unsigned less than</li>
5191 <li><tt>ule</tt>: unsigned less or equal</li>
5192 <li><tt>sgt</tt>: signed greater than</li>
5193 <li><tt>sge</tt>: signed greater or equal</li>
5194 <li><tt>slt</tt>: signed less than</li>
5195 <li><tt>sle</tt>: signed less or equal</li>
5196</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005197
Chris Lattner3b19d652007-01-15 01:54:13 +00005198<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005199 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5200 typed. They must also be identical types.</p>
5201
Reid Spencerf3a70a62006-11-18 21:50:54 +00005202<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005203<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5204 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +00005205 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005206 result, as follows:</p>
5207
Reid Spencerf3a70a62006-11-18 21:50:54 +00005208<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005209 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005210 <tt>false</tt> otherwise. No sign interpretation is necessary or
5211 performed.</li>
5212
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005213 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005214 <tt>false</tt> otherwise. No sign interpretation is necessary or
5215 performed.</li>
5216
Reid Spencerf3a70a62006-11-18 21:50:54 +00005217 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005218 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5219
Reid Spencerf3a70a62006-11-18 21:50:54 +00005220 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005221 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5222 to <tt>op2</tt>.</li>
5223
Reid Spencerf3a70a62006-11-18 21:50:54 +00005224 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005225 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5226
Reid Spencerf3a70a62006-11-18 21:50:54 +00005227 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005228 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5229
Reid Spencerf3a70a62006-11-18 21:50:54 +00005230 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005231 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5232
Reid Spencerf3a70a62006-11-18 21:50:54 +00005233 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005234 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5235 to <tt>op2</tt>.</li>
5236
Reid Spencerf3a70a62006-11-18 21:50:54 +00005237 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005238 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5239
Reid Spencerf3a70a62006-11-18 21:50:54 +00005240 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005241 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005242</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005243
Reid Spencerf3a70a62006-11-18 21:50:54 +00005244<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005245 values are compared as if they were integers.</p>
5246
5247<p>If the operands are integer vectors, then they are compared element by
5248 element. The result is an <tt>i1</tt> vector with the same number of elements
5249 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005250
5251<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005252<pre>
5253 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +00005254 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5255 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5256 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5257 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5258 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005259</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +00005260
5261<p>Note that the code generator does not yet support vector types with
5262 the <tt>icmp</tt> instruction.</p>
5263
Reid Spencerf3a70a62006-11-18 21:50:54 +00005264</div>
5265
5266<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005267<h4>
5268 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5269</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005270
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005271<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005272
Reid Spencerf3a70a62006-11-18 21:50:54 +00005273<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005274<pre>
5275 &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 +00005276</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005277
Reid Spencerf3a70a62006-11-18 21:50:54 +00005278<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005279<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5280 values based on comparison of its operands.</p>
5281
5282<p>If the operands are floating point scalars, then the result type is a boolean
Nick Lewyckyec38da42009-09-27 00:45:11 +00005283(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005284
5285<p>If the operands are floating point vectors, then the result type is a vector
5286 of boolean with the same number of elements as the operands being
5287 compared.</p>
5288
Reid Spencerf3a70a62006-11-18 21:50:54 +00005289<h5>Arguments:</h5>
5290<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005291 the condition code indicating the kind of comparison to perform. It is not a
5292 value, just a keyword. The possible condition code are:</p>
5293
Reid Spencerf3a70a62006-11-18 21:50:54 +00005294<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +00005295 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005296 <li><tt>oeq</tt>: ordered and equal</li>
5297 <li><tt>ogt</tt>: ordered and greater than </li>
5298 <li><tt>oge</tt>: ordered and greater than or equal</li>
5299 <li><tt>olt</tt>: ordered and less than </li>
5300 <li><tt>ole</tt>: ordered and less than or equal</li>
5301 <li><tt>one</tt>: ordered and not equal</li>
5302 <li><tt>ord</tt>: ordered (no nans)</li>
5303 <li><tt>ueq</tt>: unordered or equal</li>
5304 <li><tt>ugt</tt>: unordered or greater than </li>
5305 <li><tt>uge</tt>: unordered or greater than or equal</li>
5306 <li><tt>ult</tt>: unordered or less than </li>
5307 <li><tt>ule</tt>: unordered or less than or equal</li>
5308 <li><tt>une</tt>: unordered or not equal</li>
5309 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +00005310 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005311</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005312
Jeff Cohenb627eab2007-04-29 01:07:00 +00005313<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005314 <i>unordered</i> means that either operand may be a QNAN.</p>
5315
5316<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5317 a <a href="#t_floating">floating point</a> type or
5318 a <a href="#t_vector">vector</a> of floating point type. They must have
5319 identical types.</p>
5320
Reid Spencerf3a70a62006-11-18 21:50:54 +00005321<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +00005322<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005323 according to the condition code given as <tt>cond</tt>. If the operands are
5324 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +00005325 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005326 follows:</p>
5327
Reid Spencerf3a70a62006-11-18 21:50:54 +00005328<ol>
5329 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005330
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005331 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005332 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5333
Reid Spencerb7f26282006-11-19 03:00:14 +00005334 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00005335 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005336
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005337 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005338 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5339
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005340 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005341 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5342
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005343 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005344 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5345
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005346 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005347 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5348
Reid Spencerb7f26282006-11-19 03:00:14 +00005349 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005350
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005351 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005352 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5353
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005354 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005355 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5356
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005357 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005358 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5359
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005360 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005361 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5362
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005363 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005364 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5365
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005366 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005367 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5368
Reid Spencerb7f26282006-11-19 03:00:14 +00005369 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005370
Reid Spencerf3a70a62006-11-18 21:50:54 +00005371 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5372</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005373
5374<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005375<pre>
5376 &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanf72fb672008-09-09 01:02:47 +00005377 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5378 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5379 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +00005380</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +00005381
5382<p>Note that the code generator does not yet support vector types with
5383 the <tt>fcmp</tt> instruction.</p>
5384
Reid Spencerf3a70a62006-11-18 21:50:54 +00005385</div>
5386
Reid Spencer2fd21e62006-11-08 01:18:52 +00005387<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005388<h4>
Chris Lattner5568e942008-05-20 20:48:21 +00005389 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005390</h4>
Chris Lattner5568e942008-05-20 20:48:21 +00005391
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005392<div>
Chris Lattner5568e942008-05-20 20:48:21 +00005393
Reid Spencer2fd21e62006-11-08 01:18:52 +00005394<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005395<pre>
5396 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5397</pre>
Chris Lattner5568e942008-05-20 20:48:21 +00005398
Reid Spencer2fd21e62006-11-08 01:18:52 +00005399<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005400<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5401 SSA graph representing the function.</p>
5402
Reid Spencer2fd21e62006-11-08 01:18:52 +00005403<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005404<p>The type of the incoming values is specified with the first type field. After
5405 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5406 one pair for each predecessor basic block of the current block. Only values
5407 of <a href="#t_firstclass">first class</a> type may be used as the value
5408 arguments to the PHI node. Only labels may be used as the label
5409 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00005410
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005411<p>There must be no non-phi instructions between the start of a basic block and
5412 the PHI instructions: i.e. PHI instructions must be first in a basic
5413 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00005414
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005415<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5416 occur on the edge from the corresponding predecessor block to the current
5417 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5418 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +00005419
Reid Spencer2fd21e62006-11-08 01:18:52 +00005420<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005421<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005422 specified by the pair corresponding to the predecessor basic block that
5423 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +00005424
Reid Spencer2fd21e62006-11-08 01:18:52 +00005425<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +00005426<pre>
5427Loop: ; Infinite loop that counts from 0 on up...
5428 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5429 %nextindvar = add i32 %indvar, 1
5430 br label %Loop
5431</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005432
Reid Spencer2fd21e62006-11-08 01:18:52 +00005433</div>
5434
Chris Lattnercc37aae2004-03-12 05:50:16 +00005435<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005436<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005437 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005438</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005439
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005440<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005441
5442<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005443<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +00005444 &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>
5445
Dan Gohman0e451ce2008-10-14 16:51:45 +00005446 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +00005447</pre>
5448
5449<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005450<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5451 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005452
5453
5454<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005455<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5456 values indicating the condition, and two values of the
5457 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5458 vectors and the condition is a scalar, then entire vectors are selected, not
5459 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005460
5461<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005462<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5463 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005464
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005465<p>If the condition is a vector of i1, then the value arguments must be vectors
5466 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005467
5468<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005469<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +00005470 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +00005471</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +00005472
5473<p>Note that the code generator does not yet support conditions
5474 with vector type.</p>
5475
Chris Lattnercc37aae2004-03-12 05:50:16 +00005476</div>
5477
Robert Bocchino05ccd702006-01-15 20:48:27 +00005478<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005479<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +00005480 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005481</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +00005482
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005483<div>
Chris Lattner2bff5242005-05-06 05:47:36 +00005484
Chris Lattner00950542001-06-06 20:29:01 +00005485<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00005486<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +00005487 &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 +00005488</pre>
5489
Chris Lattner00950542001-06-06 20:29:01 +00005490<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00005491<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00005492
Chris Lattner00950542001-06-06 20:29:01 +00005493<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00005494<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00005495
Chris Lattner6536cfe2002-05-06 22:08:29 +00005496<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005497 <li>The optional "tail" marker indicates that the callee function does not
5498 access any allocas or varargs in the caller. Note that calls may be
5499 marked "tail" even if they do not occur before
5500 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5501 present, the function call is eligible for tail call optimization,
5502 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +00005503 optimized into a jump</a>. The code generator may optimize calls marked
5504 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5505 sibling call optimization</a> when the caller and callee have
5506 matching signatures, or 2) forced tail call optimization when the
5507 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005508 <ul>
5509 <li>Caller and callee both have the calling
5510 convention <tt>fastcc</tt>.</li>
5511 <li>The call is in tail position (ret immediately follows call and ret
5512 uses value of call or is void).</li>
5513 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +00005514 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005515 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5516 constraints are met.</a></li>
5517 </ul>
5518 </li>
Devang Patelf642f472008-10-06 18:50:38 +00005519
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005520 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5521 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +00005522 defaults to using C calling conventions. The calling convention of the
5523 call must match the calling convention of the target function, or else the
5524 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +00005525
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005526 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5527 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5528 '<tt>inreg</tt>' attributes are valid here.</li>
5529
5530 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5531 type of the return value. Functions that return no value are marked
5532 <tt><a href="#t_void">void</a></tt>.</li>
5533
5534 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5535 being invoked. The argument types must match the types implied by this
5536 signature. This type can be omitted if the function is not varargs and if
5537 the function type does not return a pointer to a function.</li>
5538
5539 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5540 be invoked. In most cases, this is a direct function invocation, but
5541 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5542 to function value.</li>
5543
5544 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +00005545 signature argument types and parameter attributes. All arguments must be
5546 of <a href="#t_firstclass">first class</a> type. If the function
5547 signature indicates the function accepts a variable number of arguments,
5548 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005549
5550 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5551 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5552 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +00005553</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +00005554
Chris Lattner00950542001-06-06 20:29:01 +00005555<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005556<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5557 a specified function, with its incoming arguments bound to the specified
5558 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5559 function, control flow continues with the instruction after the function
5560 call, and the return value of the function is bound to the result
5561 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +00005562
Chris Lattner00950542001-06-06 20:29:01 +00005563<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +00005564<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +00005565 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00005566 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +00005567 %X = tail call i32 @foo() <i>; yields i32</i>
5568 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5569 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +00005570
5571 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +00005572 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +00005573 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5574 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +00005575 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +00005576 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +00005577</pre>
5578
Dale Johannesen07de8d12009-09-24 18:38:21 +00005579<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +00005580standard C99 library as being the C99 library functions, and may perform
5581optimizations or generate code for them under that assumption. This is
5582something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00005583freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +00005584
Misha Brukman9d0919f2003-11-08 01:05:38 +00005585</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005586
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005587<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005588<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +00005589 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005590</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005591
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005592<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005593
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005594<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005595<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005596 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +00005597</pre>
5598
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005599<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005600<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005601 the "variable argument" area of a function call. It is used to implement the
5602 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005603
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005604<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005605<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5606 argument. It returns a value of the specified argument type and increments
5607 the <tt>va_list</tt> to point to the next argument. The actual type
5608 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005609
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005610<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005611<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5612 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5613 to the next argument. For more information, see the variable argument
5614 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005615
5616<p>It is legal for this instruction to be called in a function which does not
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005617 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5618 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005619
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005620<p><tt>va_arg</tt> is an LLVM instruction instead of
5621 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5622 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005623
Chris Lattner8d1a81d2003-10-18 05:51:36 +00005624<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +00005625<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5626
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005627<p>Note that the code generator does not yet fully support va_arg on many
5628 targets. Also, it does not currently support va_arg with aggregate types on
5629 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +00005630
Misha Brukman9d0919f2003-11-08 01:05:38 +00005631</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005632
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005633</div>
5634
5635</div>
5636
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005637<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005638<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +00005639<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +00005640
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005641<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005642
5643<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005644 well known names and semantics and are required to follow certain
5645 restrictions. Overall, these intrinsics represent an extension mechanism for
5646 the LLVM language that does not require changing all of the transformations
5647 in LLVM when adding to the language (or the bitcode reader/writer, the
5648 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005649
John Criswellfc6b8952005-05-16 16:17:45 +00005650<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005651 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5652 begin with this prefix. Intrinsic functions must always be external
5653 functions: you cannot define the body of intrinsic functions. Intrinsic
5654 functions may only be used in call or invoke instructions: it is illegal to
5655 take the address of an intrinsic function. Additionally, because intrinsic
5656 functions are part of the LLVM language, it is required if any are added that
5657 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005658
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005659<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5660 family of functions that perform the same operation but on different data
5661 types. Because LLVM can represent over 8 million different integer types,
5662 overloading is used commonly to allow an intrinsic function to operate on any
5663 integer type. One or more of the argument types or the result type can be
5664 overloaded to accept any integer type. Argument types may also be defined as
5665 exactly matching a previous argument's type or the result type. This allows
5666 an intrinsic function which accepts multiple arguments, but needs all of them
5667 to be of the same type, to only be overloaded with respect to a single
5668 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005669
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005670<p>Overloaded intrinsics will have the names of its overloaded argument types
5671 encoded into its function name, each preceded by a period. Only those types
5672 which are overloaded result in a name suffix. Arguments whose type is matched
5673 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5674 can take an integer of any width and returns an integer of exactly the same
5675 integer width. This leads to a family of functions such as
5676 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5677 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5678 suffix is required. Because the argument's type is matched against the return
5679 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +00005680
Eric Christopher6c7e8a02009-12-05 02:46:03 +00005681<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005682 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00005683
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005684<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005685<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00005686 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005687</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00005688
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005689<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005690
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005691<p>Variable argument support is defined in LLVM with
5692 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5693 intrinsic functions. These functions are related to the similarly named
5694 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005695
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005696<p>All of these functions operate on arguments that use a target-specific value
5697 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5698 not define what this type is, so all transformations should be prepared to
5699 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005700
Chris Lattner374ab302006-05-15 17:26:46 +00005701<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005702 instruction and the variable argument handling intrinsic functions are
5703 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005704
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00005705<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005706define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +00005707 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +00005708 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +00005709 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005710 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +00005711
5712 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +00005713 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +00005714
5715 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +00005716 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +00005717 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +00005718 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005719 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +00005720
5721 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005722 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +00005723 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +00005724}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +00005725
5726declare void @llvm.va_start(i8*)
5727declare void @llvm.va_copy(i8*, i8*)
5728declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +00005729</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +00005730
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005731<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005732<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005733 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005734</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +00005735
5736
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005737<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005738
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005739<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005740<pre>
5741 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5742</pre>
5743
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005744<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005745<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5746 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005747
5748<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +00005749<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005750
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005751<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +00005752<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005753 macro available in C. In a target-dependent way, it initializes
5754 the <tt>va_list</tt> element to which the argument points, so that the next
5755 call to <tt>va_arg</tt> will produce the first variable argument passed to
5756 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5757 need to know the last argument of the function as the compiler can figure
5758 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005759
Misha Brukman9d0919f2003-11-08 01:05:38 +00005760</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005761
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005762<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005763<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005764 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005765</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +00005766
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005767<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005768
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005769<h5>Syntax:</h5>
5770<pre>
5771 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5772</pre>
5773
5774<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005775<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005776 which has been initialized previously
5777 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5778 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005779
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005780<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005781<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005782
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005783<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +00005784<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005785 macro available in C. In a target-dependent way, it destroys
5786 the <tt>va_list</tt> element to which the argument points. Calls
5787 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5788 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5789 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +00005790
Misha Brukman9d0919f2003-11-08 01:05:38 +00005791</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005792
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005793<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005794<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005795 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005796</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +00005797
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005798<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005799
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005800<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005801<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00005802 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +00005803</pre>
5804
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005805<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005806<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005807 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005808
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005809<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +00005810<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005811 The second argument is a pointer to a <tt>va_list</tt> element to copy
5812 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005813
Chris Lattnerd9ad5b32003-05-08 04:57:36 +00005814<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +00005815<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005816 macro available in C. In a target-dependent way, it copies the
5817 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5818 element. This intrinsic is necessary because
5819 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5820 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005821
Misha Brukman9d0919f2003-11-08 01:05:38 +00005822</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00005823
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005824</div>
5825
Chris Lattner33aec9e2004-02-12 17:01:32 +00005826<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005827<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +00005828 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005829</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +00005830
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005831<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005832
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005833<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +00005834Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005835intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5836roots on the stack</a>, as well as garbage collector implementations that
5837require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5838barriers. Front-ends for type-safe garbage collected languages should generate
5839these intrinsics to make use of the LLVM garbage collectors. For more details,
5840see <a href="GarbageCollection.html">Accurate Garbage Collection with
5841LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +00005842
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005843<p>The garbage collection intrinsics only operate on objects in the generic
5844 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +00005845
Chris Lattnerd7923912004-05-23 21:06:01 +00005846<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005847<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005848 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005849</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +00005850
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005851<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005852
5853<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005854<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00005855 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +00005856</pre>
5857
5858<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +00005859<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005860 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005861
5862<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005863<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005864 root pointer. The second pointer (which must be either a constant or a
5865 global value address) contains the meta-data to be associated with the
5866 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005867
5868<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +00005869<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005870 location. At compile-time, the code generator generates information to allow
5871 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5872 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5873 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005874
5875</div>
5876
Chris Lattnerd7923912004-05-23 21:06:01 +00005877<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005878<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005879 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005880</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +00005881
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005882<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005883
5884<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005885<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00005886 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +00005887</pre>
5888
5889<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005890<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005891 locations, allowing garbage collector implementations that require read
5892 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005893
5894<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +00005895<p>The second argument is the address to read from, which should be an address
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005896 allocated from the garbage collector. The first object is a pointer to the
5897 start of the referenced object, if needed by the language runtime (otherwise
5898 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005899
5900<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005901<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005902 instruction, but may be replaced with substantially more complex code by the
5903 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5904 may only be used in a function which <a href="#gc">specifies a GC
5905 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005906
5907</div>
5908
Chris Lattnerd7923912004-05-23 21:06:01 +00005909<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005910<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005911 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005912</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +00005913
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005914<div>
Chris Lattnerd7923912004-05-23 21:06:01 +00005915
5916<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005917<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00005918 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +00005919</pre>
5920
5921<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005922<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005923 locations, allowing garbage collector implementations that require write
5924 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005925
5926<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +00005927<p>The first argument is the reference to store, the second is the start of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005928 object to store it to, and the third is the address of the field of Obj to
5929 store to. If the runtime does not require a pointer to the object, Obj may
5930 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005931
5932<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +00005933<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005934 instruction, but may be replaced with substantially more complex code by the
5935 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5936 may only be used in a function which <a href="#gc">specifies a GC
5937 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +00005938
5939</div>
5940
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005941</div>
5942
Chris Lattnerd7923912004-05-23 21:06:01 +00005943<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005944<h3>
Chris Lattner10610642004-02-14 04:08:35 +00005945 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005946</h3>
Chris Lattner10610642004-02-14 04:08:35 +00005947
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005948<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005949
5950<p>These intrinsics are provided by LLVM to expose special features that may
5951 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005952
Chris Lattner10610642004-02-14 04:08:35 +00005953<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005954<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005955 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005956</h4>
Chris Lattner10610642004-02-14 04:08:35 +00005957
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005958<div>
Chris Lattner10610642004-02-14 04:08:35 +00005959
5960<h5>Syntax:</h5>
5961<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +00005962 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00005963</pre>
5964
5965<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005966<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5967 target-specific value indicating the return address of the current function
5968 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005969
5970<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005971<p>The argument to this intrinsic indicates which function to return the address
5972 for. Zero indicates the calling function, one indicates its caller, etc.
5973 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005974
5975<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005976<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5977 indicating the return address of the specified call frame, or zero if it
5978 cannot be identified. The value returned by this intrinsic is likely to be
5979 incorrect or 0 for arguments other than zero, so it should only be used for
5980 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005981
Bill Wendlinge910b4c2009-07-20 02:29:24 +00005982<p>Note that calling this intrinsic does not prevent function inlining or other
5983 aggressive transformations, so the value returned may not be that of the
5984 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +00005985
Chris Lattner10610642004-02-14 04:08:35 +00005986</div>
5987
Chris Lattner10610642004-02-14 04:08:35 +00005988<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005989<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00005990 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00005991</h4>
Chris Lattner10610642004-02-14 04:08:35 +00005992
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00005993<div>
Chris Lattner10610642004-02-14 04:08:35 +00005994
5995<h5>Syntax:</h5>
5996<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00005997 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00005998</pre>
5999
6000<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006001<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6002 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006003
6004<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006005<p>The argument to this intrinsic indicates which function to return the frame
6006 pointer for. Zero indicates the calling function, one indicates its caller,
6007 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006008
6009<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006010<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6011 indicating the frame address of the specified call frame, or zero if it
6012 cannot be identified. The value returned by this intrinsic is likely to be
6013 incorrect or 0 for arguments other than zero, so it should only be used for
6014 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006015
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006016<p>Note that calling this intrinsic does not prevent function inlining or other
6017 aggressive transformations, so the value returned may not be that of the
6018 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006019
Chris Lattner10610642004-02-14 04:08:35 +00006020</div>
6021
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006022<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006023<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006024 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006025</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +00006026
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006027<div>
Chris Lattner57e1f392006-01-13 02:03:13 +00006028
6029<h5>Syntax:</h5>
6030<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006031 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +00006032</pre>
6033
6034<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006035<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6036 of the function stack, for use
6037 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6038 useful for implementing language features like scoped automatic variable
6039 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006040
6041<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006042<p>This intrinsic returns a opaque pointer value that can be passed
6043 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6044 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6045 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6046 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6047 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6048 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006049
6050</div>
6051
6052<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006053<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006054 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006055</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +00006056
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006057<div>
Chris Lattner57e1f392006-01-13 02:03:13 +00006058
6059<h5>Syntax:</h5>
6060<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006061 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +00006062</pre>
6063
6064<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006065<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6066 the function stack to the state it was in when the
6067 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6068 executed. This is useful for implementing language features like scoped
6069 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006070
6071<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006072<p>See the description
6073 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +00006074
6075</div>
6076
Chris Lattner57e1f392006-01-13 02:03:13 +00006077<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006078<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006079 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006080</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006081
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006082<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006083
6084<h5>Syntax:</h5>
6085<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +00006086 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 +00006087</pre>
6088
6089<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006090<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6091 insert a prefetch instruction if supported; otherwise, it is a noop.
6092 Prefetches have no effect on the behavior of the program but can change its
6093 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006094
6095<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006096<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6097 specifier determining if the fetch should be for a read (0) or write (1),
6098 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +00006099 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6100 specifies whether the prefetch is performed on the data (1) or instruction (0)
6101 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6102 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006103
6104<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006105<p>This intrinsic does not modify the behavior of the program. In particular,
6106 prefetches cannot trap and do not produce a value. On targets that support
6107 this intrinsic, the prefetch can provide hints to the processor cache for
6108 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +00006109
6110</div>
6111
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006112<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006113<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006114 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006115</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006116
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006117<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006118
6119<h5>Syntax:</h5>
6120<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +00006121 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006122</pre>
6123
6124<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006125<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6126 Counter (PC) in a region of code to simulators and other tools. The method
6127 is target specific, but it is expected that the marker will use exported
6128 symbols to transmit the PC of the marker. The marker makes no guarantees
6129 that it will remain with any specific instruction after optimizations. It is
6130 possible that the presence of a marker will inhibit optimizations. The
6131 intended use is to be inserted after optimizations to allow correlations of
6132 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006133
6134<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006135<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006136
6137<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006138<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006139 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +00006140
6141</div>
6142
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006143<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006144<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006145 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006146</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006147
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006148<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006149
6150<h5>Syntax:</h5>
6151<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00006152 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006153</pre>
6154
6155<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006156<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6157 counter register (or similar low latency, high accuracy clocks) on those
6158 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6159 should map to RPCC. As the backing counters overflow quickly (on the order
6160 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006161
6162<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006163<p>When directly supported, reading the cycle counter should not modify any
6164 memory. Implementations are allowed to either return a application specific
6165 value or a system wide value. On backends without support, this is lowered
6166 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +00006167
6168</div>
6169
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006170</div>
6171
Chris Lattner10610642004-02-14 04:08:35 +00006172<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006173<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006174 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006175</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006176
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006177<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006178
6179<p>LLVM provides intrinsics for a few important standard C library functions.
6180 These intrinsics allow source-language front-ends to pass information about
6181 the alignment of the pointer arguments to the code generator, providing
6182 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006183
Chris Lattner33aec9e2004-02-12 17:01:32 +00006184<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006185<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006186 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006187</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006188
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006189<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006190
6191<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006192<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
Mon P Wange88909b2010-04-07 06:35:53 +00006193 integer bit width and for different address spaces. Not all targets support
6194 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006195
Chris Lattner33aec9e2004-02-12 17:01:32 +00006196<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006197 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006198 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006199 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006200 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +00006201</pre>
6202
6203<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006204<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6205 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006206
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006207<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +00006208 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6209 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006210
6211<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006212
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006213<p>The first argument is a pointer to the destination, the second is a pointer
6214 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +00006215 number of bytes to copy, the fourth argument is the alignment of the
6216 source and destination locations, and the fifth is a boolean indicating a
6217 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006218
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006219<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006220 then the caller guarantees that both the source and destination pointers are
6221 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +00006222
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00006223<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6224 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6225 The detailed access behavior is not very cleanly specified and it is unwise
6226 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +00006227
Chris Lattner33aec9e2004-02-12 17:01:32 +00006228<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006229
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006230<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6231 source location to the destination location, which are not allowed to
6232 overlap. It copies "len" bytes of memory over. If the argument is known to
6233 be aligned to some boundary, this can be specified as the fourth argument,
6234 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +00006235
Chris Lattner33aec9e2004-02-12 17:01:32 +00006236</div>
6237
Chris Lattner0eb51b42004-02-12 18:10:10 +00006238<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006239<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006240 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006241</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006242
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006243<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006244
6245<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +00006246<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Mon P Wange88909b2010-04-07 06:35:53 +00006247 width and for different address space. Not all targets support all bit
6248 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006249
Chris Lattner0eb51b42004-02-12 18:10:10 +00006250<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006251 declare void @llvm.memmove.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006252 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006253 declare void @llvm.memmove.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +00006254 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +00006255</pre>
6256
6257<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006258<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6259 source location to the destination location. It is similar to the
6260 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6261 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006262
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006263<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +00006264 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6265 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006266
6267<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006268
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006269<p>The first argument is a pointer to the destination, the second is a pointer
6270 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +00006271 number of bytes to copy, the fourth argument is the alignment of the
6272 source and destination locations, and the fifth is a boolean indicating a
6273 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006274
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006275<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006276 then the caller guarantees that the source and destination pointers are
6277 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +00006278
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00006279<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6280 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6281 The detailed access behavior is not very cleanly specified and it is unwise
6282 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +00006283
Chris Lattner0eb51b42004-02-12 18:10:10 +00006284<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +00006285
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006286<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6287 source location to the destination location, which may overlap. It copies
6288 "len" bytes of memory over. If the argument is known to be aligned to some
6289 boundary, this can be specified as the fourth argument, otherwise it should
6290 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +00006291
Chris Lattner0eb51b42004-02-12 18:10:10 +00006292</div>
6293
Chris Lattner10610642004-02-14 04:08:35 +00006294<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006295<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006296 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006297</h4>
Chris Lattner10610642004-02-14 04:08:35 +00006298
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006299<div>
Chris Lattner10610642004-02-14 04:08:35 +00006300
6301<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +00006302<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
John Criswellcdcbbfc2010-07-30 16:30:28 +00006303 width and for different address spaces. However, not all targets support all
6304 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006305
Chris Lattner10610642004-02-14 04:08:35 +00006306<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006307 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +00006308 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +00006309 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +00006310 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +00006311</pre>
6312
6313<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006314<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6315 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006316
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006317<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +00006318 intrinsic does not return a value and takes extra alignment/volatile
6319 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006320
6321<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006322<p>The first argument is a pointer to the destination to fill, the second is the
John Criswellcdcbbfc2010-07-30 16:30:28 +00006323 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006324 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +00006325 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006326
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00006327<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006328 then the caller guarantees that the destination pointer is aligned to that
6329 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006330
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +00006331<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6332 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6333 The detailed access behavior is not very cleanly specified and it is unwise
6334 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +00006335
Chris Lattner10610642004-02-14 04:08:35 +00006336<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006337<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6338 at the destination location. If the argument is known to be aligned to some
6339 boundary, this can be specified as the fourth argument, otherwise it should
6340 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +00006341
Chris Lattner10610642004-02-14 04:08:35 +00006342</div>
6343
Chris Lattner32006282004-06-11 02:28:03 +00006344<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006345<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006346 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006347</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +00006348
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006349<div>
Chris Lattnera4d74142005-07-21 01:29:16 +00006350
6351<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006352<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6353 floating point or vector of floating point type. Not all targets support all
6354 types however.</p>
6355
Chris Lattnera4d74142005-07-21 01:29:16 +00006356<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +00006357 declare float @llvm.sqrt.f32(float %Val)
6358 declare double @llvm.sqrt.f64(double %Val)
6359 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6360 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6361 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +00006362</pre>
6363
6364<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006365<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6366 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6367 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6368 behavior for negative numbers other than -0.0 (which allows for better
6369 optimization, because there is no need to worry about errno being
6370 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +00006371
6372<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006373<p>The argument and return value are floating point numbers of the same
6374 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +00006375
6376<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006377<p>This function returns the sqrt of the specified operand if it is a
6378 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +00006379
Chris Lattnera4d74142005-07-21 01:29:16 +00006380</div>
6381
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006382<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006383<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006384 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006385</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006386
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006387<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006388
6389<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006390<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6391 floating point or vector of floating point type. Not all targets support all
6392 types however.</p>
6393
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006394<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +00006395 declare float @llvm.powi.f32(float %Val, i32 %power)
6396 declare double @llvm.powi.f64(double %Val, i32 %power)
6397 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6398 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6399 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006400</pre>
6401
6402<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006403<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6404 specified (positive or negative) power. The order of evaluation of
6405 multiplications is not defined. When a vector of floating point type is
6406 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006407
6408<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006409<p>The second argument is an integer power, and the first is a value to raise to
6410 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006411
6412<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006413<p>This function returns the first value raised to the second power with an
6414 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006415
Chris Lattnerf4d252d2006-09-08 06:34:02 +00006416</div>
6417
Dan Gohman91c284c2007-10-15 20:30:11 +00006418<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006419<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006420 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006421</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006422
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006423<div>
Dan Gohman91c284c2007-10-15 20:30:11 +00006424
6425<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006426<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6427 floating point or vector of floating point type. Not all targets support all
6428 types however.</p>
6429
Dan Gohman91c284c2007-10-15 20:30:11 +00006430<pre>
6431 declare float @llvm.sin.f32(float %Val)
6432 declare double @llvm.sin.f64(double %Val)
6433 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6434 declare fp128 @llvm.sin.f128(fp128 %Val)
6435 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6436</pre>
6437
6438<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006439<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006440
6441<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006442<p>The argument and return value are floating point numbers of the same
6443 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006444
6445<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006446<p>This function returns the sine of the specified operand, returning the same
6447 values as the libm <tt>sin</tt> functions would, and handles error conditions
6448 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006449
Dan Gohman91c284c2007-10-15 20:30:11 +00006450</div>
6451
6452<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006453<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006454 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006455</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006456
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006457<div>
Dan Gohman91c284c2007-10-15 20:30:11 +00006458
6459<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006460<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6461 floating point or vector of floating point type. Not all targets support all
6462 types however.</p>
6463
Dan Gohman91c284c2007-10-15 20:30:11 +00006464<pre>
6465 declare float @llvm.cos.f32(float %Val)
6466 declare double @llvm.cos.f64(double %Val)
6467 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6468 declare fp128 @llvm.cos.f128(fp128 %Val)
6469 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6470</pre>
6471
6472<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006473<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006474
6475<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006476<p>The argument and return value are floating point numbers of the same
6477 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006478
6479<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006480<p>This function returns the cosine of the specified operand, returning the same
6481 values as the libm <tt>cos</tt> functions would, and handles error conditions
6482 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006483
Dan Gohman91c284c2007-10-15 20:30:11 +00006484</div>
6485
6486<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006487<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006488 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006489</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +00006490
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006491<div>
Dan Gohman91c284c2007-10-15 20:30:11 +00006492
6493<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006494<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6495 floating point or vector of floating point type. Not all targets support all
6496 types however.</p>
6497
Dan Gohman91c284c2007-10-15 20:30:11 +00006498<pre>
6499 declare float @llvm.pow.f32(float %Val, float %Power)
6500 declare double @llvm.pow.f64(double %Val, double %Power)
6501 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6502 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6503 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6504</pre>
6505
6506<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006507<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6508 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006509
6510<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006511<p>The second argument is a floating point power, and the first is a value to
6512 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006513
6514<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006515<p>This function returns the first value raised to the second power, returning
6516 the same values as the libm <tt>pow</tt> functions would, and handles error
6517 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +00006518
Dan Gohman91c284c2007-10-15 20:30:11 +00006519</div>
6520
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006521</div>
6522
Dan Gohman4e9011c2011-05-23 21:13:03 +00006523<!-- _______________________________________________________________________ -->
6524<h4>
6525 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
6526</h4>
6527
6528<div>
6529
6530<h5>Syntax:</h5>
6531<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
6532 floating point or vector of floating point type. Not all targets support all
6533 types however.</p>
6534
6535<pre>
6536 declare float @llvm.exp.f32(float %Val)
6537 declare double @llvm.exp.f64(double %Val)
6538 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
6539 declare fp128 @llvm.exp.f128(fp128 %Val)
6540 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
6541</pre>
6542
6543<h5>Overview:</h5>
6544<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
6545
6546<h5>Arguments:</h5>
6547<p>The argument and return value are floating point numbers of the same
6548 type.</p>
6549
6550<h5>Semantics:</h5>
6551<p>This function returns the same values as the libm <tt>exp</tt> functions
6552 would, and handles error conditions in the same way.</p>
6553
6554</div>
6555
6556<!-- _______________________________________________________________________ -->
6557<h4>
6558 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
6559</h4>
6560
6561<div>
6562
6563<h5>Syntax:</h5>
6564<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
6565 floating point or vector of floating point type. Not all targets support all
6566 types however.</p>
6567
6568<pre>
6569 declare float @llvm.log.f32(float %Val)
6570 declare double @llvm.log.f64(double %Val)
6571 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
6572 declare fp128 @llvm.log.f128(fp128 %Val)
6573 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
6574</pre>
6575
6576<h5>Overview:</h5>
6577<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
6578
6579<h5>Arguments:</h5>
6580<p>The argument and return value are floating point numbers of the same
6581 type.</p>
6582
6583<h5>Semantics:</h5>
6584<p>This function returns the same values as the libm <tt>log</tt> functions
6585 would, and handles error conditions in the same way.</p>
6586
Cameron Zwarich33390842011-07-08 21:39:21 +00006587<h4>
6588 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
6589</h4>
6590
6591<div>
6592
6593<h5>Syntax:</h5>
6594<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
6595 floating point or vector of floating point type. Not all targets support all
6596 types however.</p>
6597
6598<pre>
6599 declare float @llvm.fma.f32(float %a, float %b, float %c)
6600 declare double @llvm.fma.f64(double %a, double %b, double %c)
6601 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
6602 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
6603 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
6604</pre>
6605
6606<h5>Overview:</h5>
Cameron Zwarichabc43e62011-07-08 22:13:55 +00006607<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarich33390842011-07-08 21:39:21 +00006608 operation.</p>
6609
6610<h5>Arguments:</h5>
6611<p>The argument and return value are floating point numbers of the same
6612 type.</p>
6613
6614<h5>Semantics:</h5>
6615<p>This function returns the same values as the libm <tt>fma</tt> functions
6616 would.</p>
6617
Dan Gohman4e9011c2011-05-23 21:13:03 +00006618</div>
6619
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006620<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006621<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +00006622 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006623</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006624
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006625<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006626
6627<p>LLVM provides intrinsics for a few important bit manipulation operations.
6628 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006629
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006630<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006631<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +00006632 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006633</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +00006634
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006635<div>
Nate Begeman7e36c472006-01-13 23:26:38 +00006636
6637<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00006638<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006639 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6640
Nate Begeman7e36c472006-01-13 23:26:38 +00006641<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00006642 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6643 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6644 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +00006645</pre>
6646
6647<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006648<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6649 values with an even number of bytes (positive multiple of 16 bits). These
6650 are useful for performing operations on data that is not in the target's
6651 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +00006652
6653<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006654<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6655 and low byte of the input i16 swapped. Similarly,
6656 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6657 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6658 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6659 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6660 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6661 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +00006662
6663</div>
6664
6665<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006666<h4>
Reid Spencer0b118202006-01-16 21:12:35 +00006667 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006668</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006669
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006670<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006671
6672<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +00006673<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Owen Andersonf1ac4652011-07-01 21:52:38 +00006674 width, or on any vector with integer elements. Not all targets support all
6675 bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006676
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006677<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006678 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006679 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00006680 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006681 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6682 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +00006683 declare &lt;2 x i32&gt; @llvm.ctpop.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006684</pre>
6685
6686<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006687<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6688 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006689
6690<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006691<p>The only argument is the value to be counted. The argument may be of any
Owen Andersonf1ac4652011-07-01 21:52:38 +00006692 integer type, or a vector with integer elements.
6693 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006694
6695<h5>Semantics:</h5>
Owen Andersonf1ac4652011-07-01 21:52:38 +00006696<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
6697 element of a vector.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006698
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006699</div>
6700
6701<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006702<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +00006703 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006704</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006705
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006706<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006707
6708<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006709<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Owen Andersonf1ac4652011-07-01 21:52:38 +00006710 integer bit width, or any vector whose elements are integers. Not all
6711 targets support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006712
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006713<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00006714 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6715 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00006716 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006717 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6718 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +00006719 declare &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src;gt)
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006720</pre>
6721
6722<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006723<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6724 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006725
6726<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006727<p>The only argument is the value to be counted. The argument may be of any
Owen Andersonf1ac4652011-07-01 21:52:38 +00006728 integer type, or any vector type with integer element type.
6729 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006730
6731<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006732<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +00006733 zeros in a variable, or within each element of the vector if the operation
6734 is of vector type. If the src == 0 then the result is the size in bits of
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006735 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006736
Andrew Lenharthec370fd2005-05-03 18:01:48 +00006737</div>
Chris Lattner32006282004-06-11 02:28:03 +00006738
Chris Lattnereff29ab2005-05-15 19:39:26 +00006739<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006740<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +00006741 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006742</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006743
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006744<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006745
6746<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006747<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Owen Andersonf1ac4652011-07-01 21:52:38 +00006748 integer bit width, or any vector of integer elements. Not all targets
6749 support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006750
Chris Lattnereff29ab2005-05-15 19:39:26 +00006751<pre>
Chandler Carruth69940402007-08-04 01:51:18 +00006752 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6753 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +00006754 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +00006755 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6756 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +00006757 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +00006758</pre>
6759
6760<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006761<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6762 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006763
6764<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006765<p>The only argument is the value to be counted. The argument may be of any
Owen Andersonf1ac4652011-07-01 21:52:38 +00006766 integer type, or a vectory with integer element type.. The return type
6767 must match the argument type.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006768
6769<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006770<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +00006771 zeros in a variable, or within each element of a vector.
6772 If the src == 0 then the result is the size in bits of
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006773 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +00006774
Chris Lattnereff29ab2005-05-15 19:39:26 +00006775</div>
6776
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006777</div>
6778
Bill Wendlingda01af72009-02-08 04:04:40 +00006779<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006780<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +00006781 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006782</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +00006783
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006784<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006785
6786<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +00006787
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006788<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006789<h4>
6790 <a name="int_sadd_overflow">
6791 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
6792 </a>
6793</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006794
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006795<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006796
6797<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006798<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006799 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006800
6801<pre>
6802 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6803 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6804 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6805</pre>
6806
6807<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006808<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006809 a signed addition of the two arguments, and indicate whether an overflow
6810 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006811
6812<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006813<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006814 be of integer types of any bit width, but they must have the same bit
6815 width. The second element of the result structure must be of
6816 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6817 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006818
6819<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006820<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006821 a signed addition of the two variables. They return a structure &mdash; the
6822 first element of which is the signed summation, and the second element of
6823 which is a bit specifying if the signed summation resulted in an
6824 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006825
6826<h5>Examples:</h5>
6827<pre>
6828 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6829 %sum = extractvalue {i32, i1} %res, 0
6830 %obit = extractvalue {i32, i1} %res, 1
6831 br i1 %obit, label %overflow, label %normal
6832</pre>
6833
6834</div>
6835
6836<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006837<h4>
6838 <a name="int_uadd_overflow">
6839 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
6840 </a>
6841</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006842
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006843<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006844
6845<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006846<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006847 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006848
6849<pre>
6850 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6851 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6852 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6853</pre>
6854
6855<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006856<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006857 an unsigned addition of the two arguments, and indicate whether a carry
6858 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006859
6860<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006861<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006862 be of integer types of any bit width, but they must have the same bit
6863 width. The second element of the result structure must be of
6864 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6865 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006866
6867<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006868<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006869 an unsigned addition of the two arguments. They return a structure &mdash;
6870 the first element of which is the sum, and the second element of which is a
6871 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006872
6873<h5>Examples:</h5>
6874<pre>
6875 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6876 %sum = extractvalue {i32, i1} %res, 0
6877 %obit = extractvalue {i32, i1} %res, 1
6878 br i1 %obit, label %carry, label %normal
6879</pre>
6880
6881</div>
6882
6883<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006884<h4>
6885 <a name="int_ssub_overflow">
6886 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
6887 </a>
6888</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006889
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006890<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006891
6892<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006893<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006894 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006895
6896<pre>
6897 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6898 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6899 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6900</pre>
6901
6902<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006903<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006904 a signed subtraction of the two arguments, and indicate whether an overflow
6905 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006906
6907<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006908<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006909 be of integer types of any bit width, but they must have the same bit
6910 width. The second element of the result structure must be of
6911 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6912 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006913
6914<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006915<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006916 a signed subtraction of the two arguments. They return a structure &mdash;
6917 the first element of which is the subtraction, and the second element of
6918 which is a bit specifying if the signed subtraction resulted in an
6919 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006920
6921<h5>Examples:</h5>
6922<pre>
6923 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6924 %sum = extractvalue {i32, i1} %res, 0
6925 %obit = extractvalue {i32, i1} %res, 1
6926 br i1 %obit, label %overflow, label %normal
6927</pre>
6928
6929</div>
6930
6931<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006932<h4>
6933 <a name="int_usub_overflow">
6934 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
6935 </a>
6936</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006937
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006938<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006939
6940<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006941<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006942 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006943
6944<pre>
6945 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6946 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6947 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6948</pre>
6949
6950<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006951<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006952 an unsigned subtraction of the two arguments, and indicate whether an
6953 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006954
6955<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006956<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006957 be of integer types of any bit width, but they must have the same bit
6958 width. The second element of the result structure must be of
6959 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6960 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006961
6962<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006963<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006964 an unsigned subtraction of the two arguments. They return a structure &mdash;
6965 the first element of which is the subtraction, and the second element of
6966 which is a bit specifying if the unsigned subtraction resulted in an
6967 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006968
6969<h5>Examples:</h5>
6970<pre>
6971 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6972 %sum = extractvalue {i32, i1} %res, 0
6973 %obit = extractvalue {i32, i1} %res, 1
6974 br i1 %obit, label %overflow, label %normal
6975</pre>
6976
6977</div>
6978
6979<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00006980<h4>
6981 <a name="int_smul_overflow">
6982 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
6983 </a>
6984</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006985
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00006986<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006987
6988<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006989<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00006990 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00006991
6992<pre>
6993 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6994 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6995 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6996</pre>
6997
6998<h5>Overview:</h5>
6999
7000<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007001 a signed multiplication of the two arguments, and indicate whether an
7002 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00007003
7004<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00007005<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007006 be of integer types of any bit width, but they must have the same bit
7007 width. The second element of the result structure must be of
7008 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7009 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00007010
7011<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00007012<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007013 a signed multiplication of the two arguments. They return a structure &mdash;
7014 the first element of which is the multiplication, and the second element of
7015 which is a bit specifying if the signed multiplication resulted in an
7016 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +00007017
7018<h5>Examples:</h5>
7019<pre>
7020 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7021 %sum = extractvalue {i32, i1} %res, 0
7022 %obit = extractvalue {i32, i1} %res, 1
7023 br i1 %obit, label %overflow, label %normal
7024</pre>
7025
Reid Spencerf86037f2007-04-11 23:23:49 +00007026</div>
7027
Bill Wendling41b485c2009-02-08 23:00:09 +00007028<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007029<h4>
7030 <a name="int_umul_overflow">
7031 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7032 </a>
7033</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +00007034
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007035<div>
Bill Wendling41b485c2009-02-08 23:00:09 +00007036
7037<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00007038<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007039 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00007040
7041<pre>
7042 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7043 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7044 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7045</pre>
7046
7047<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00007048<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007049 a unsigned multiplication of the two arguments, and indicate whether an
7050 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00007051
7052<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00007053<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007054 be of integer types of any bit width, but they must have the same bit
7055 width. The second element of the result structure must be of
7056 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7057 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00007058
7059<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +00007060<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007061 an unsigned multiplication of the two arguments. They return a structure
7062 &mdash; the first element of which is the multiplication, and the second
7063 element of which is a bit specifying if the unsigned multiplication resulted
7064 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +00007065
7066<h5>Examples:</h5>
7067<pre>
7068 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7069 %sum = extractvalue {i32, i1} %res, 0
7070 %obit = extractvalue {i32, i1} %res, 1
7071 br i1 %obit, label %overflow, label %normal
7072</pre>
7073
7074</div>
7075
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007076</div>
7077
Chris Lattner8ff75902004-01-06 05:31:32 +00007078<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007079<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007080 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007081</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007082
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007083<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007084
Chris Lattner0cec9c82010-03-15 04:12:21 +00007085<p>Half precision floating point is a storage-only format. This means that it is
7086 a dense encoding (in memory) but does not support computation in the
7087 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +00007088
Chris Lattner0cec9c82010-03-15 04:12:21 +00007089<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +00007090 value as an i16, then convert it to float with <a
7091 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7092 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +00007093 double etc). To store the value back to memory, it is first converted to
7094 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +00007095 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7096 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007097
7098<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007099<h4>
7100 <a name="int_convert_to_fp16">
7101 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7102 </a>
7103</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007104
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007105<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007106
7107<h5>Syntax:</h5>
7108<pre>
7109 declare i16 @llvm.convert.to.fp16(f32 %a)
7110</pre>
7111
7112<h5>Overview:</h5>
7113<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7114 a conversion from single precision floating point format to half precision
7115 floating point format.</p>
7116
7117<h5>Arguments:</h5>
7118<p>The intrinsic function contains single argument - the value to be
7119 converted.</p>
7120
7121<h5>Semantics:</h5>
7122<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7123 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +00007124 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +00007125 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007126
7127<h5>Examples:</h5>
7128<pre>
7129 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7130 store i16 %res, i16* @x, align 2
7131</pre>
7132
7133</div>
7134
7135<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007136<h4>
7137 <a name="int_convert_from_fp16">
7138 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7139 </a>
7140</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007141
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007142<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007143
7144<h5>Syntax:</h5>
7145<pre>
7146 declare f32 @llvm.convert.from.fp16(i16 %a)
7147</pre>
7148
7149<h5>Overview:</h5>
7150<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7151 a conversion from half precision floating point format to single precision
7152 floating point format.</p>
7153
7154<h5>Arguments:</h5>
7155<p>The intrinsic function contains single argument - the value to be
7156 converted.</p>
7157
7158<h5>Semantics:</h5>
7159<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +00007160 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +00007161 precision floating point format. The input half-float value is represented by
7162 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007163
7164<h5>Examples:</h5>
7165<pre>
7166 %a = load i16* @x, align 2
7167 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7168</pre>
7169
7170</div>
7171
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007172</div>
7173
Anton Korobeynikovf02e7302010-03-14 18:42:47 +00007174<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007175<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00007176 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007177</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +00007178
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007179<div>
Chris Lattner8ff75902004-01-06 05:31:32 +00007180
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007181<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7182 prefix), are described in
7183 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7184 Level Debugging</a> document.</p>
7185
7186</div>
Chris Lattner8ff75902004-01-06 05:31:32 +00007187
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007188<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007189<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007190 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007191</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007192
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007193<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007194
7195<p>The LLVM exception handling intrinsics (which all start with
7196 <tt>llvm.eh.</tt> prefix), are described in
7197 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7198 Handling</a> document.</p>
7199
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007200</div>
7201
Tanya Lattner6d806e92007-06-15 20:50:54 +00007202<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007203<h3>
Duncan Sandsf7331b32007-09-11 14:10:23 +00007204 <a name="int_trampoline">Trampoline Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007205</h3>
Duncan Sands36397f52007-07-27 12:58:54 +00007206
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007207<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007208
7209<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +00007210 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7211 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007212 function pointer lacking the nest parameter - the caller does not need to
7213 provide a value for it. Instead, the value to use is stored in advance in a
7214 "trampoline", a block of memory usually allocated on the stack, which also
7215 contains code to splice the nest value into the argument list. This is used
7216 to implement the GCC nested function address extension.</p>
7217
7218<p>For example, if the function is
7219 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7220 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7221 follows:</p>
7222
Benjamin Kramer26fe25f2010-07-13 12:26:09 +00007223<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +00007224 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7225 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007226 %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 +00007227 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +00007228</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007229
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007230<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7231 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007232
Duncan Sands36397f52007-07-27 12:58:54 +00007233<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007234<h4>
7235 <a name="int_it">
7236 '<tt>llvm.init.trampoline</tt>' Intrinsic
7237 </a>
7238</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007239
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007240<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007241
Duncan Sands36397f52007-07-27 12:58:54 +00007242<h5>Syntax:</h5>
7243<pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007244 declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands36397f52007-07-27 12:58:54 +00007245</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007246
Duncan Sands36397f52007-07-27 12:58:54 +00007247<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007248<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
7249 function pointer suitable for executing it.</p>
7250
Duncan Sands36397f52007-07-27 12:58:54 +00007251<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007252<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7253 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7254 sufficiently aligned block of memory; this memory is written to by the
7255 intrinsic. Note that the size and the alignment are target-specific - LLVM
7256 currently provides no portable way of determining them, so a front-end that
7257 generates this intrinsic needs to have some target-specific knowledge.
7258 The <tt>func</tt> argument must hold a function bitcast to
7259 an <tt>i8*</tt>.</p>
7260
Duncan Sands36397f52007-07-27 12:58:54 +00007261<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007262<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7263 dependent code, turning it into a function. A pointer to this function is
7264 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7265 function pointer type</a> before being called. The new function's signature
7266 is the same as that of <tt>func</tt> with any arguments marked with
7267 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7268 is allowed, and it must be of pointer type. Calling the new function is
7269 equivalent to calling <tt>func</tt> with the same argument list, but
7270 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7271 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7272 by <tt>tramp</tt> is modified, then the effect of any later call to the
7273 returned function pointer is undefined.</p>
7274
Duncan Sands36397f52007-07-27 12:58:54 +00007275</div>
7276
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007277</div>
7278
Duncan Sands36397f52007-07-27 12:58:54 +00007279<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007280<h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007281 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007282</h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007283
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007284<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007285
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007286<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7287 hardware constructs for atomic operations and memory synchronization. This
7288 provides an interface to the hardware, not an interface to the programmer. It
7289 is aimed at a low enough level to allow any programming models or APIs
7290 (Application Programming Interfaces) which need atomic behaviors to map
7291 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7292 hardware provides a "universal IR" for source languages, it also provides a
7293 starting point for developing a "universal" atomic operation and
7294 synchronization IR.</p>
7295
7296<p>These do <em>not</em> form an API such as high-level threading libraries,
7297 software transaction memory systems, atomic primitives, and intrinsic
7298 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7299 application libraries. The hardware interface provided by LLVM should allow
7300 a clean implementation of all of these APIs and parallel programming models.
7301 No one model or paradigm should be selected above others unless the hardware
7302 itself ubiquitously does so.</p>
7303
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007304<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007305<h4>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007306 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007307</h4>
7308
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007309<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007310<h5>Syntax:</h5>
7311<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007312 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 +00007313</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007314
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007315<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007316<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7317 specific pairs of memory access types.</p>
7318
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007319<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007320<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7321 The first four arguments enables a specific barrier as listed below. The
Dan Gohmanb55a1ee2010-03-01 17:41:39 +00007322 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007323 memory.</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007324
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007325<ul>
7326 <li><tt>ll</tt>: load-load barrier</li>
7327 <li><tt>ls</tt>: load-store barrier</li>
7328 <li><tt>sl</tt>: store-load barrier</li>
7329 <li><tt>ss</tt>: store-store barrier</li>
7330 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7331</ul>
7332
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007333<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007334<p>This intrinsic causes the system to enforce some ordering constraints upon
7335 the loads and stores of the program. This barrier does not
7336 indicate <em>when</em> any events will occur, it only enforces
7337 an <em>order</em> in which they occur. For any of the specified pairs of load
7338 and store operations (f.ex. load-load, or store-load), all of the first
7339 operations preceding the barrier will complete before any of the second
7340 operations succeeding the barrier begin. Specifically the semantics for each
7341 pairing is as follows:</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007342
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007343<ul>
7344 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7345 after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007346 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007347 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007348 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007349 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007350 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007351 load after the barrier begins.</li>
7352</ul>
7353
7354<p>These semantics are applied with a logical "and" behavior when more than one
7355 is enabled in a single memory barrier intrinsic.</p>
7356
7357<p>Backends may implement stronger barriers than those requested when they do
7358 not support as fine grained a barrier as requested. Some architectures do
7359 not need all types of barriers and on such architectures, these become
7360 noops.</p>
7361
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007362<h5>Example:</h5>
7363<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007364%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7365%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007366 store i32 4, %ptr
7367
7368%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Evan Cheng0b0669a2011-06-29 17:14:00 +00007369 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false, i1 true)
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007370 <i>; guarantee the above finishes</i>
7371 store i32 8, %ptr <i>; before this begins</i>
7372</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007373
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007374</div>
7375
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007376<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007377<h4>
Mon P Wang28873102008-06-25 08:15:39 +00007378 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007379</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007380
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007381<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007382
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007383<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007384<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7385 any integer bit width and for different address spaces. Not all targets
7386 support all bit widths however.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007387
7388<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007389 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7390 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7391 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7392 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 +00007393</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007394
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007395<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007396<p>This loads a value in memory and compares it to a given value. If they are
7397 equal, it stores a new value into the memory.</p>
7398
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007399<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007400<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7401 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7402 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7403 this integer type. While any bit width integer may be used, targets may only
7404 lower representations they support in hardware.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007405
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007406<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007407<p>This entire intrinsic must be executed atomically. It first loads the value
7408 in memory pointed to by <tt>ptr</tt> and compares it with the
7409 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7410 memory. The loaded value is yielded in all cases. This provides the
7411 equivalent of an atomic compare-and-swap operation within the SSA
7412 framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007413
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007414<h5>Examples:</h5>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007415<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007416%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7417%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007418 store i32 4, %ptr
7419
7420%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007421%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007422 <i>; yields {i32}:result1 = 4</i>
7423%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7424%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7425
7426%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007427%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007428 <i>; yields {i32}:result2 = 8</i>
7429%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7430
7431%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7432</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007433
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007434</div>
7435
7436<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007437<h4>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007438 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007439</h4>
7440
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007441<div>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007442<h5>Syntax:</h5>
7443
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007444<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7445 integer bit width. Not all targets support all bit widths however.</p>
7446
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007447<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007448 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7449 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7450 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7451 declare i64 @llvm.atomic.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;val&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007452</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007453
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007454<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007455<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7456 the value from memory. It then stores the value in <tt>val</tt> in the memory
7457 at <tt>ptr</tt>.</p>
7458
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007459<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007460<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7461 the <tt>val</tt> argument and the result must be integers of the same bit
7462 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7463 integer type. The targets may only lower integer representations they
7464 support.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007465
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007466<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007467<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7468 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7469 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007470
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007471<h5>Examples:</h5>
7472<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007473%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7474%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007475 store i32 4, %ptr
7476
7477%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007478%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007479 <i>; yields {i32}:result1 = 4</i>
7480%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7481%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7482
7483%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007484%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007485 <i>; yields {i32}:result2 = 8</i>
7486
7487%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7488%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7489</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007490
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007491</div>
7492
7493<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007494<h4>
Mon P Wang28873102008-06-25 08:15:39 +00007495 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007496</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007497
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007498<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007499
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007500<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007501<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7502 any integer bit width. Not all targets support all bit widths however.</p>
7503
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007504<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007505 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7506 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7507 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7508 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007509</pre>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007510
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007511<h5>Overview:</h5>
7512<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7513 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7514
7515<h5>Arguments:</h5>
7516<p>The intrinsic takes two arguments, the first a pointer to an integer value
7517 and the second an integer value. The result is also an integer value. These
7518 integer types can have any bit width, but they must all have the same bit
7519 width. The targets may only lower integer representations they support.</p>
7520
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007521<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007522<p>This intrinsic does a series of operations atomically. It first loads the
7523 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7524 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007525
7526<h5>Examples:</h5>
7527<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007528%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7529%ptr = bitcast i8* %mallocP to i32*
7530 store i32 4, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007531%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007532 <i>; yields {i32}:result1 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007533%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007534 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007535%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007536 <i>; yields {i32}:result3 = 10</i>
Mon P Wang28873102008-06-25 08:15:39 +00007537%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007538</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007539
Andrew Lenharthab0b9492008-02-21 06:45:13 +00007540</div>
7541
Mon P Wang28873102008-06-25 08:15:39 +00007542<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007543<h4>
Mon P Wang28873102008-06-25 08:15:39 +00007544 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007545</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007546
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007547<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007548
Mon P Wang28873102008-06-25 08:15:39 +00007549<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007550<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7551 any integer bit width and for different address spaces. Not all targets
7552 support all bit widths however.</p>
7553
Mon P Wang28873102008-06-25 08:15:39 +00007554<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007555 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7556 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7557 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7558 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007559</pre>
Mon P Wang28873102008-06-25 08:15:39 +00007560
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007561<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007562<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007563 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7564
7565<h5>Arguments:</h5>
7566<p>The intrinsic takes two arguments, the first a pointer to an integer value
7567 and the second an integer value. The result is also an integer value. These
7568 integer types can have any bit width, but they must all have the same bit
7569 width. The targets may only lower integer representations they support.</p>
7570
Mon P Wang28873102008-06-25 08:15:39 +00007571<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007572<p>This intrinsic does a series of operations atomically. It first loads the
7573 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7574 result to <tt>ptr</tt>. It yields the original value stored
7575 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007576
7577<h5>Examples:</h5>
7578<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007579%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7580%ptr = bitcast i8* %mallocP to i32*
7581 store i32 8, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007582%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang28873102008-06-25 08:15:39 +00007583 <i>; yields {i32}:result1 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007584%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang28873102008-06-25 08:15:39 +00007585 <i>; yields {i32}:result2 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007586%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang28873102008-06-25 08:15:39 +00007587 <i>; yields {i32}:result3 = 2</i>
7588%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7589</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007590
Mon P Wang28873102008-06-25 08:15:39 +00007591</div>
7592
7593<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007594<h4>
7595 <a name="int_atomic_load_and">
7596 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
7597 </a>
7598 <br>
7599 <a name="int_atomic_load_nand">
7600 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
7601 </a>
7602 <br>
7603 <a name="int_atomic_load_or">
7604 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
7605 </a>
7606 <br>
7607 <a name="int_atomic_load_xor">
7608 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
7609 </a>
7610</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007611
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007612<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007613
Mon P Wang28873102008-06-25 08:15:39 +00007614<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007615<p>These are overloaded intrinsics. You can
7616 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7617 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7618 bit width and for different address spaces. Not all targets support all bit
7619 widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007620
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007621<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007622 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7623 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7624 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7625 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007626</pre>
7627
7628<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007629 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7630 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7631 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7632 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007633</pre>
7634
7635<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007636 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7637 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7638 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7639 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007640</pre>
7641
7642<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007643 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7644 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7645 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7646 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007647</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007648
Mon P Wang28873102008-06-25 08:15:39 +00007649<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007650<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7651 the value stored in memory at <tt>ptr</tt>. It yields the original value
7652 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007653
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007654<h5>Arguments:</h5>
7655<p>These intrinsics take two arguments, the first a pointer to an integer value
7656 and the second an integer value. The result is also an integer value. These
7657 integer types can have any bit width, but they must all have the same bit
7658 width. The targets may only lower integer representations they support.</p>
7659
Mon P Wang28873102008-06-25 08:15:39 +00007660<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007661<p>These intrinsics does a series of operations atomically. They first load the
7662 value stored at <tt>ptr</tt>. They then do the bitwise
7663 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7664 original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007665
7666<h5>Examples:</h5>
7667<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007668%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7669%ptr = bitcast i8* %mallocP to i32*
7670 store i32 0x0F0F, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007671%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +00007672 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007673%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +00007674 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007675%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +00007676 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007677%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +00007678 <i>; yields {i32}:result3 = FF</i>
7679%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7680</pre>
Mon P Wang28873102008-06-25 08:15:39 +00007681
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007682</div>
Mon P Wang28873102008-06-25 08:15:39 +00007683
7684<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007685<h4>
7686 <a name="int_atomic_load_max">
7687 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
7688 </a>
7689 <br>
7690 <a name="int_atomic_load_min">
7691 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
7692 </a>
7693 <br>
7694 <a name="int_atomic_load_umax">
7695 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
7696 </a>
7697 <br>
7698 <a name="int_atomic_load_umin">
7699 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
7700 </a>
7701</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007702
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007703<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007704
Mon P Wang28873102008-06-25 08:15:39 +00007705<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007706<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7707 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7708 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7709 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007710
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007711<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007712 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7713 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7714 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7715 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007716</pre>
7717
7718<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007719 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7720 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7721 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7722 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007723</pre>
7724
7725<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007726 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7727 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7728 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7729 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007730</pre>
7731
7732<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007733 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7734 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7735 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7736 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +00007737</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007738
Mon P Wang28873102008-06-25 08:15:39 +00007739<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +00007740<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007741 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7742 original value at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007743
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007744<h5>Arguments:</h5>
7745<p>These intrinsics take two arguments, the first a pointer to an integer value
7746 and the second an integer value. The result is also an integer value. These
7747 integer types can have any bit width, but they must all have the same bit
7748 width. The targets may only lower integer representations they support.</p>
7749
Mon P Wang28873102008-06-25 08:15:39 +00007750<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007751<p>These intrinsics does a series of operations atomically. They first load the
7752 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7753 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7754 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +00007755
7756<h5>Examples:</h5>
7757<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +00007758%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7759%ptr = bitcast i8* %mallocP to i32*
7760 store i32 7, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007761%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang28873102008-06-25 08:15:39 +00007762 <i>; yields {i32}:result0 = 7</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007763%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang28873102008-06-25 08:15:39 +00007764 <i>; yields {i32}:result1 = -2</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007765%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang28873102008-06-25 08:15:39 +00007766 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007767%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang28873102008-06-25 08:15:39 +00007768 <i>; yields {i32}:result3 = 8</i>
7769%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7770</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007771
Mon P Wang28873102008-06-25 08:15:39 +00007772</div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007773
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007774</div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007775
7776<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007777<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +00007778 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007779</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +00007780
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007781<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007782
7783<p>This class of intrinsics exists to information about the lifetime of memory
7784 objects and ranges where variables are immutable.</p>
7785
Nick Lewyckycc271862009-10-13 07:03:23 +00007786<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007787<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007788 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007789</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007790
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007791<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007792
7793<h5>Syntax:</h5>
7794<pre>
7795 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7796</pre>
7797
7798<h5>Overview:</h5>
7799<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7800 object's lifetime.</p>
7801
7802<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +00007803<p>The first argument is a constant integer representing the size of the
7804 object, or -1 if it is variable sized. The second argument is a pointer to
7805 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007806
7807<h5>Semantics:</h5>
7808<p>This intrinsic indicates that before this point in the code, the value of the
7809 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
Nick Lewycky8d336592009-10-27 16:56:58 +00007810 never be used and has an undefined value. A load from the pointer that
7811 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +00007812 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7813
7814</div>
7815
7816<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007817<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007818 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007819</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007820
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007821<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007822
7823<h5>Syntax:</h5>
7824<pre>
7825 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7826</pre>
7827
7828<h5>Overview:</h5>
7829<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7830 object's lifetime.</p>
7831
7832<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +00007833<p>The first argument is a constant integer representing the size of the
7834 object, or -1 if it is variable sized. The second argument is a pointer to
7835 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007836
7837<h5>Semantics:</h5>
7838<p>This intrinsic indicates that after this point in the code, the value of the
7839 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7840 never be used and has an undefined value. Any stores into the memory object
7841 following this intrinsic may be removed as dead.
7842
7843</div>
7844
7845<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007846<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007847 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007848</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007849
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007850<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007851
7852<h5>Syntax:</h5>
7853<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +00007854 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +00007855</pre>
7856
7857<h5>Overview:</h5>
7858<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7859 a memory object will not change.</p>
7860
7861<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +00007862<p>The first argument is a constant integer representing the size of the
7863 object, or -1 if it is variable sized. The second argument is a pointer to
7864 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007865
7866<h5>Semantics:</h5>
7867<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7868 the return value, the referenced memory location is constant and
7869 unchanging.</p>
7870
7871</div>
7872
7873<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007874<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007875 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007876</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +00007877
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007878<div>
Nick Lewyckycc271862009-10-13 07:03:23 +00007879
7880<h5>Syntax:</h5>
7881<pre>
7882 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7883</pre>
7884
7885<h5>Overview:</h5>
7886<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7887 a memory object are mutable.</p>
7888
7889<h5>Arguments:</h5>
7890<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +00007891 The second argument is a constant integer representing the size of the
7892 object, or -1 if it is variable sized and the third argument is a pointer
7893 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +00007894
7895<h5>Semantics:</h5>
7896<p>This intrinsic indicates that the memory is mutable again.</p>
7897
7898</div>
7899
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007900</div>
7901
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +00007902<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007903<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007904 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007905</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007906
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007907<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007908
7909<p>This class of intrinsics is designed to be generic and has no specific
7910 purpose.</p>
7911
Tanya Lattner6d806e92007-06-15 20:50:54 +00007912<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007913<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007914 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007915</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007916
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007917<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007918
7919<h5>Syntax:</h5>
7920<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007921 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 +00007922</pre>
7923
7924<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007925<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007926
7927<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007928<p>The first argument is a pointer to a value, the second is a pointer to a
7929 global string, the third is a pointer to a global string which is the source
7930 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007931
7932<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007933<p>This intrinsic allows annotation of local variables with arbitrary strings.
7934 This can be useful for special purpose optimizations that want to look for
7935 these annotations. These have no other defined use, they are ignored by code
7936 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +00007937
Tanya Lattner6d806e92007-06-15 20:50:54 +00007938</div>
7939
Tanya Lattnerb6367882007-09-21 22:59:12 +00007940<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007941<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +00007942 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007943</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007944
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007945<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007946
7947<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007948<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7949 any integer bit width.</p>
7950
Tanya Lattnerb6367882007-09-21 22:59:12 +00007951<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00007952 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7953 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7954 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7955 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7956 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 +00007957</pre>
7958
7959<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007960<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007961
7962<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007963<p>The first argument is an integer value (result of some expression), the
7964 second is a pointer to a global string, the third is a pointer to a global
7965 string which is the source file name, and the last argument is the line
7966 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007967
7968<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007969<p>This intrinsic allows annotations to be put on arbitrary expressions with
7970 arbitrary strings. This can be useful for special purpose optimizations that
7971 want to look for these annotations. These have no other defined use, they
7972 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +00007973
Tanya Lattnerb6367882007-09-21 22:59:12 +00007974</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +00007975
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007976<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007977<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007978 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00007979</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007980
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00007981<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007982
7983<h5>Syntax:</h5>
7984<pre>
7985 declare void @llvm.trap()
7986</pre>
7987
7988<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007989<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007990
7991<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007992<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007993
7994<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00007995<p>This intrinsics is lowered to the target dependent trap instruction. If the
7996 target does not have a trap instruction, this intrinsic will be lowered to
7997 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007998
Anton Korobeynikov4cb86182008-01-15 22:31:34 +00007999</div>
8000
Bill Wendling69e4adb2008-11-19 05:56:17 +00008001<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00008002<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +00008003 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00008004</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00008005
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00008006<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00008007
Bill Wendling69e4adb2008-11-19 05:56:17 +00008008<h5>Syntax:</h5>
8009<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00008010 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +00008011</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00008012
Bill Wendling69e4adb2008-11-19 05:56:17 +00008013<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00008014<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8015 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8016 ensure that it is placed on the stack before local variables.</p>
8017
Bill Wendling69e4adb2008-11-19 05:56:17 +00008018<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00008019<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8020 arguments. The first argument is the value loaded from the stack
8021 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8022 that has enough space to hold the value of the guard.</p>
8023
Bill Wendling69e4adb2008-11-19 05:56:17 +00008024<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +00008025<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8026 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8027 stack. This is to ensure that if a local variable on the stack is
8028 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +00008029 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +00008030 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8031 function.</p>
8032
Bill Wendling69e4adb2008-11-19 05:56:17 +00008033</div>
8034
Eric Christopher0e671492009-11-30 08:03:53 +00008035<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00008036<h4>
Eric Christopher0e671492009-11-30 08:03:53 +00008037 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +00008038</h4>
Eric Christopher0e671492009-11-30 08:03:53 +00008039
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00008040<div>
Eric Christopher0e671492009-11-30 08:03:53 +00008041
8042<h5>Syntax:</h5>
8043<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +00008044 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8045 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +00008046</pre>
8047
8048<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +00008049<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8050 the optimizers to determine at compile time whether a) an operation (like
8051 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8052 runtime check for overflow isn't necessary. An object in this context means
8053 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +00008054
8055<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +00008056<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +00008057 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +00008058 is a boolean 0 or 1. This argument determines whether you want the
8059 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher8295a0a2009-12-23 00:29:49 +00008060 1, variables are not allowed.</p>
8061
Eric Christopher0e671492009-11-30 08:03:53 +00008062<h5>Semantics:</h5>
8063<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +00008064 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8065 depending on the <tt>type</tt> argument, if the size cannot be determined at
8066 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +00008067
8068</div>
8069
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +00008070</div>
8071
8072</div>
8073
Chris Lattner00950542001-06-06 20:29:01 +00008074<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +00008075<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008076<address>
8077 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +00008078 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008079 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +00008080 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008081
8082 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +00008083 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +00008084 Last modified: $Date$
8085</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +00008086
Misha Brukman9d0919f2003-11-08 01:05:38 +00008087</body>
8088</html>