Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 44579680111b807613703ab401db3b8c0148e36c / . / docs / LangRef.html
blob: ab9f1caae69638b0f99af0ed75ce51126583d11e [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]14
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]15<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]25 <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 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]28 <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 +0000[diff] [blame]29 <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 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]37 <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 +0000[diff] [blame]41 </ol>
42 </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]43 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]44 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]45 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]46 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]47 <li><a href="#aliasstructure">Aliases</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]49 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]50 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]51 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]53 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]56 </ol>
57 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]58 <li><a href="#typesystem">Type System</a>
59 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]60 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]61 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]62 <ol>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]63 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]64 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]65 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]66 <li><a href="#t_void">Void Type</a></li>
67 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]68 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]69 </ol>
70 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]71 <li><a href="#t_derived">Derived Types</a>
72 <ol>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]73 <li><a href="#t_aggregate">Aggregate Types</a>
74 <ol>
75 <li><a href="#t_array">Array Type</a></li>
76 <li><a href="#t_struct">Structure Type</a></li>
77 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]78 <li><a href="#t_vector">Vector Type</a></li>
79 </ol>
80 </li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]81 <li><a href="#t_function">Function Type</a></li>
82 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]83 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]84 </ol>
85 </li>
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]86 <li><a href="#t_uprefs">Type Up-references</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]87 </ol>
88 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]89 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]90 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]91 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]92 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]93 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
94 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]95 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattnerf9d078e2009-10-27 21:19:13 +0000[diff] [blame]96 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]97 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]98 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]99 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]104 </ol>
105 </li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]106 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
107 <ol>
108 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]109 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
110 Global Variable</a></li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]111 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
112 Global Variable</a></li>
113 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
114 Global Variable</a></li>
115 </ol>
116 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]117 <li><a href="#instref">Instruction Reference</a>
118 <ol>
119 <li><a href="#terminators">Terminator Instructions</a>
120 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]121 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
122 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]123 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]124 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]125 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]126 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]127 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]128 </ol>
129 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]130 <li><a href="#binaryops">Binary Operations</a>
131 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]132 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]133 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]134 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]135 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]136 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]137 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]138 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
139 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
140 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]141 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
142 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
143 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]144 </ol>
145 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]146 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
147 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]148 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
149 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
150 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]151 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]152 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]153 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]154 </ol>
155 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]156 <li><a href="#vectorops">Vector Operations</a>
157 <ol>
158 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
159 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
160 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]161 </ol>
162 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]163 <li><a href="#aggregateops">Aggregate Operations</a>
164 <ol>
165 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
166 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
167 </ol>
168 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]169 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]170 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]171 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]172 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
173 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
174 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]175 </ol>
176 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]177 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]178 <ol>
179 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
180 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
181 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
182 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
183 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]184 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
185 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
186 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
187 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]188 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
189 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]190 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]191 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]192 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]193 <li><a href="#otherops">Other Operations</a>
194 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]195 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
196 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]197 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]198 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]199 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]200 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]201 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]202 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]203 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]204 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]205 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]206 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]207 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
208 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]209 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
210 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
211 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]212 </ol>
213 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]214 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
215 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]216 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
217 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
218 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]219 </ol>
220 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]221 <li><a href="#int_codegen">Code Generator Intrinsics</a>
222 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]223 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
224 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
225 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
226 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
227 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
228 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000[diff] [blame]229 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]230 </ol>
231 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]232 <li><a href="#int_libc">Standard C Library Intrinsics</a>
233 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]234 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
235 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
238 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]239 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]242 </ol>
243 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]244 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]245 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]246 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]247 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
248 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
249 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]250 </ol>
251 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]252 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
253 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]254 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
255 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
256 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
257 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
258 <li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]259 <li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]260 </ol>
261 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]262 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
263 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]264 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
265 <li><a href="#int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a></li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]266 </ol>
267 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]268 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]269 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]270 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]271 <ol>
272 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]273 </ol>
274 </li>
Bill Wendling3c44f5b2008-11-18 22:10:53 +0000[diff] [blame]275 <li><a href="#int_atomics">Atomic intrinsics</a>
276 <ol>
277 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
278 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
279 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
280 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
281 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
282 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
283 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
284 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
285 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
286 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
287 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
288 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
289 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
290 </ol>
291 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]292 <li><a href="#int_memorymarkers">Memory Use Markers</a>
293 <ol>
294 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
295 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
296 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
297 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
298 </ol>
299 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]300 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]301 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]302 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]303 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]304 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]305 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]306 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]307 '<tt>llvm.trap</tt>' Intrinsic</a></li>
308 <li><a href="#int_stackprotector">
309 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]310 <li><a href="#int_objectsize">
311 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]312 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]313 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]314 </ol>
315 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]316</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]317
318<div class="doc_author">
319 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
320 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]321</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]322
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]323<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]324<div class="doc_section"> <a name="abstract">Abstract </a></div>
325<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]326
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]327<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]328
329<p>This document is a reference manual for the LLVM assembly language. LLVM is
330 a Static Single Assignment (SSA) based representation that provides type
331 safety, low-level operations, flexibility, and the capability of representing
332 'all' high-level languages cleanly. It is the common code representation
333 used throughout all phases of the LLVM compilation strategy.</p>
334
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]335</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]336
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]337<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]338<div class="doc_section"> <a name="introduction">Introduction</a> </div>
339<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]340
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]341<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]342
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]343<p>The LLVM code representation is designed to be used in three different forms:
344 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
345 for fast loading by a Just-In-Time compiler), and as a human readable
346 assembly language representation. This allows LLVM to provide a powerful
347 intermediate representation for efficient compiler transformations and
348 analysis, while providing a natural means to debug and visualize the
349 transformations. The three different forms of LLVM are all equivalent. This
350 document describes the human readable representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]351
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]352<p>The LLVM representation aims to be light-weight and low-level while being
353 expressive, typed, and extensible at the same time. It aims to be a
354 "universal IR" of sorts, by being at a low enough level that high-level ideas
355 may be cleanly mapped to it (similar to how microprocessors are "universal
356 IR's", allowing many source languages to be mapped to them). By providing
357 type information, LLVM can be used as the target of optimizations: for
358 example, through pointer analysis, it can be proven that a C automatic
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]359 variable is never accessed outside of the current function, allowing it to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]360 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]361
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]362</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]363
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]364<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]365<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]366
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]367<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]368
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]369<p>It is important to note that this document describes 'well formed' LLVM
370 assembly language. There is a difference between what the parser accepts and
371 what is considered 'well formed'. For example, the following instruction is
372 syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]373
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]374<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]375%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]376</pre>
377
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]378<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
379 LLVM infrastructure provides a verification pass that may be used to verify
380 that an LLVM module is well formed. This pass is automatically run by the
381 parser after parsing input assembly and by the optimizer before it outputs
382 bitcode. The violations pointed out by the verifier pass indicate bugs in
383 transformation passes or input to the parser.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]384
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]385</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]386
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame]387<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]388
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]389<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]390<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]391<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]392
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]393<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]394
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]395<p>LLVM identifiers come in two basic types: global and local. Global
396 identifiers (functions, global variables) begin with the <tt>'@'</tt>
397 character. Local identifiers (register names, types) begin with
398 the <tt>'%'</tt> character. Additionally, there are three different formats
399 for identifiers, for different purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]400
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]401<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]402 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]403 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
404 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
405 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
406 other characters in their names can be surrounded with quotes. Special
407 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
408 ASCII code for the character in hexadecimal. In this way, any character
409 can be used in a name value, even quotes themselves.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]410
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]411 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]412 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]413
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]414 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]415 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]416</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]417
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]418<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]419 don't need to worry about name clashes with reserved words, and the set of
420 reserved words may be expanded in the future without penalty. Additionally,
421 unnamed identifiers allow a compiler to quickly come up with a temporary
422 variable without having to avoid symbol table conflicts.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]423
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]424<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]425 languages. There are keywords for different opcodes
426 ('<tt><a href="#i_add">add</a></tt>',
427 '<tt><a href="#i_bitcast">bitcast</a></tt>',
428 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
429 ('<tt><a href="#t_void">void</a></tt>',
430 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
431 reserved words cannot conflict with variable names, because none of them
432 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]433
434<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]435 '<tt>%X</tt>' by 8:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]436
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]437<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]438
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]439<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]440%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]441</pre>
442
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]443<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]444
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]445<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]446%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]447</pre>
448
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]449<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]450
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]451<pre class="doc_code">
Gabor Greifec58f752009-10-28 13:05:07 +0000[diff] [blame]452%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
453%1 = <a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]454%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]455</pre>
456
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]457<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
458 lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]459
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]460<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]461 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]462 line.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]463
464 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]465 assigned to a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]466
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]467 <li>Unnamed temporaries are numbered sequentially</li>
468</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]469
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]470<p>It also shows a convention that we follow in this document. When
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]471 demonstrating instructions, we will follow an instruction with a comment that
472 defines the type and name of value produced. Comments are shown in italic
473 text.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]474
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]475</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]476
477<!-- *********************************************************************** -->
478<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
479<!-- *********************************************************************** -->
480
481<!-- ======================================================================= -->
482<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
483</div>
484
485<div class="doc_text">
486
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]487<p>LLVM programs are composed of "Module"s, each of which is a translation unit
488 of the input programs. Each module consists of functions, global variables,
489 and symbol table entries. Modules may be combined together with the LLVM
490 linker, which merges function (and global variable) definitions, resolves
491 forward declarations, and merges symbol table entries. Here is an example of
492 the "hello world" module:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]493
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]494<pre class="doc_code">
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]495<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckydb9cd762011-01-29 01:09:53 +0000[diff] [blame]496<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 +0000[diff] [blame]497
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]498<i>; External declaration of the puts function</i>&nbsp;
499<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]500
501<i>; Definition of main function</i>
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]502define i32 @main() { <i>; i32()* </i>&nbsp;
503 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
504 %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 +0000[diff] [blame]505
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]506 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
507 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
508 <a href="#i_ret">ret</a> i32 0&nbsp;
509}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]510
511<i>; Named metadata</i>
512!1 = metadata !{i32 41}
513!foo = !{!1, null}
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]514</pre>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]515
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]516<p>This example is made up of a <a href="#globalvars">global variable</a> named
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]517 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]518 a <a href="#functionstructure">function definition</a> for
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]519 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
520 "<tt>foo"</tt>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]521
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]522<p>In general, a module is made up of a list of global values, where both
523 functions and global variables are global values. Global values are
524 represented by a pointer to a memory location (in this case, a pointer to an
525 array of char, and a pointer to a function), and have one of the
526 following <a href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]527
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]528</div>
529
530<!-- ======================================================================= -->
531<div class="doc_subsection">
532 <a name="linkage">Linkage Types</a>
533</div>
534
535<div class="doc_text">
536
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]537<p>All Global Variables and Functions have one of the following types of
538 linkage:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]539
540<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]541 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]542 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
543 by objects in the current module. In particular, linking code into a
544 module with an private global value may cause the private to be renamed as
545 necessary to avoid collisions. Because the symbol is private to the
546 module, all references can be updated. This doesn't show up in any symbol
547 table in the object file.</dd>
Rafael Espindolabb46f522009-01-15 20:18:42 +0000[diff] [blame]548
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]549 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]550 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
551 assembler and evaluated by the linker. Unlike normal strong symbols, they
552 are removed by the linker from the final linked image (executable or
553 dynamic library).</dd>
554
555 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
556 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
557 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
558 linker. The symbols are removed by the linker from the final linked image
559 (executable or dynamic library).</dd>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]560
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]561 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
562 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
563 of the object is not taken. For instance, functions that had an inline
564 definition, but the compiler decided not to inline it. Note,
565 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
566 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
567 visibility. The symbols are removed by the linker from the final linked
568 image (executable or dynamic library).</dd>
569
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]570 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling07d31772010-06-29 22:34:52 +0000[diff] [blame]571 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]572 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
573 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]574
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]575 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]576 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]577 into the object file corresponding to the LLVM module. They exist to
578 allow inlining and other optimizations to take place given knowledge of
579 the definition of the global, which is known to be somewhere outside the
580 module. Globals with <tt>available_externally</tt> linkage are allowed to
581 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
582 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]583
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]584 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]585 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner873187c2010-01-09 19:15:14 +0000[diff] [blame]586 the same name when linkage occurs. This can be used to implement
587 some forms of inline functions, templates, or other code which must be
588 generated in each translation unit that uses it, but where the body may
589 be overridden with a more definitive definition later. Unreferenced
590 <tt>linkonce</tt> globals are allowed to be discarded. Note that
591 <tt>linkonce</tt> linkage does not actually allow the optimizer to
592 inline the body of this function into callers because it doesn't know if
593 this definition of the function is the definitive definition within the
594 program or whether it will be overridden by a stronger definition.
595 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
596 linkage.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]597
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]598 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]599 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
600 <tt>linkonce</tt> linkage, except that unreferenced globals with
601 <tt>weak</tt> linkage may not be discarded. This is used for globals that
602 are declared "weak" in C source code.</dd>
603
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]604 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]605 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
606 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
607 global scope.
608 Symbols with "<tt>common</tt>" linkage are merged in the same way as
609 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]610 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]611 must have a zero initializer, and may not be marked '<a
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]612 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
613 have common linkage.</dd>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]614
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]615
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]616 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]617 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]618 pointer to array type. When two global variables with appending linkage
619 are linked together, the two global arrays are appended together. This is
620 the LLVM, typesafe, equivalent of having the system linker append together
621 "sections" with identical names when .o files are linked.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]622
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]623 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]624 <dd>The semantics of this linkage follow the ELF object file model: the symbol
625 is weak until linked, if not linked, the symbol becomes null instead of
626 being an undefined reference.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]627
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]628 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
629 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]630 <dd>Some languages allow differing globals to be merged, such as two functions
631 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]632 that only equivalent globals are ever merged (the "one definition rule"
633 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]634 and <tt>weak_odr</tt> linkage types to indicate that the global will only
635 be merged with equivalent globals. These linkage types are otherwise the
636 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]637
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]638 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]639 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]640 visible, meaning that it participates in linkage and can be used to
641 resolve external symbol references.</dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]642</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]643
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]644<p>The next two types of linkage are targeted for Microsoft Windows platform
645 only. They are designed to support importing (exporting) symbols from (to)
646 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]647
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]648<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]649 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]650 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]651 or variable via a global pointer to a pointer that is set up by the DLL
652 exporting the symbol. On Microsoft Windows targets, the pointer name is
653 formed by combining <code>__imp_</code> and the function or variable
654 name.</dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]655
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]656 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]657 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]658 pointer to a pointer in a DLL, so that it can be referenced with the
659 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
660 name is formed by combining <code>__imp_</code> and the function or
661 variable name.</dd>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]662</dl>
663
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]664<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
665 another module defined a "<tt>.LC0</tt>" variable and was linked with this
666 one, one of the two would be renamed, preventing a collision. Since
667 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
668 declarations), they are accessible outside of the current module.</p>
669
670<p>It is illegal for a function <i>declaration</i> to have any linkage type
671 other than "externally visible", <tt>dllimport</tt>
672 or <tt>extern_weak</tt>.</p>
673
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]674<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]675 or <tt>weak_odr</tt> linkages.</p>
676
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]677</div>
678
679<!-- ======================================================================= -->
680<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]681 <a name="callingconv">Calling Conventions</a>
682</div>
683
684<div class="doc_text">
685
686<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]687 and <a href="#i_invoke">invokes</a> can all have an optional calling
688 convention specified for the call. The calling convention of any pair of
689 dynamic caller/callee must match, or the behavior of the program is
690 undefined. The following calling conventions are supported by LLVM, and more
691 may be added in the future:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]692
693<dl>
694 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]695 <dd>This calling convention (the default if no other calling convention is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]696 specified) matches the target C calling conventions. This calling
697 convention supports varargs function calls and tolerates some mismatch in
698 the declared prototype and implemented declaration of the function (as
699 does normal C).</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]700
701 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]702 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]703 (e.g. by passing things in registers). This calling convention allows the
704 target to use whatever tricks it wants to produce fast code for the
705 target, without having to conform to an externally specified ABI
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]706 (Application Binary Interface).
707 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]708 when this or the GHC convention is used.</a> This calling convention
709 does not support varargs and requires the prototype of all callees to
710 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]711
712 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]713 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]714 as possible under the assumption that the call is not commonly executed.
715 As such, these calls often preserve all registers so that the call does
716 not break any live ranges in the caller side. This calling convention
717 does not support varargs and requires the prototype of all callees to
718 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]719
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]720 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
721 <dd>This calling convention has been implemented specifically for use by the
722 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
723 It passes everything in registers, going to extremes to achieve this by
724 disabling callee save registers. This calling convention should not be
725 used lightly but only for specific situations such as an alternative to
726 the <em>register pinning</em> performance technique often used when
727 implementing functional programming languages.At the moment only X86
728 supports this convention and it has the following limitations:
729 <ul>
730 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
731 floating point types are supported.</li>
732 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
733 6 floating point parameters.</li>
734 </ul>
735 This calling convention supports
736 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
737 requires both the caller and callee are using it.
738 </dd>
739
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]740 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]741 <dd>Any calling convention may be specified by number, allowing
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]742 target-specific calling conventions to be used. Target specific calling
743 conventions start at 64.</dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]744</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]745
746<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]747 support Pascal conventions or any other well-known target-independent
748 convention.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]749
750</div>
751
752<!-- ======================================================================= -->
753<div class="doc_subsection">
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]754 <a name="visibility">Visibility Styles</a>
755</div>
756
757<div class="doc_text">
758
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]759<p>All Global Variables and Functions have one of the following visibility
760 styles:</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]761
762<dl>
763 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]764 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]765 that the declaration is visible to other modules and, in shared libraries,
766 means that the declared entity may be overridden. On Darwin, default
767 visibility means that the declaration is visible to other modules. Default
768 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]769
770 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]771 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]772 object if they are in the same shared object. Usually, hidden visibility
773 indicates that the symbol will not be placed into the dynamic symbol
774 table, so no other module (executable or shared library) can reference it
775 directly.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]776
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]777 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]778 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]779 the dynamic symbol table, but that references within the defining module
780 will bind to the local symbol. That is, the symbol cannot be overridden by
781 another module.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]782</dl>
783
784</div>
785
786<!-- ======================================================================= -->
787<div class="doc_subsection">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]788 <a name="namedtypes">Named Types</a>
789</div>
790
791<div class="doc_text">
792
793<p>LLVM IR allows you to specify name aliases for certain types. This can make
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]794 it easier to read the IR and make the IR more condensed (particularly when
795 recursive types are involved). An example of a name specification is:</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]796
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]797<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]798%mytype = type { %mytype*, i32 }
799</pre>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]800
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]801<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattnerdc65f222010-08-17 23:26:04 +0000[diff] [blame]802 "<a href="#t_void">void</a>". Type name aliases may be used anywhere a type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]803 is expected with the syntax "%mytype".</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]804
805<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]806 and that you can therefore specify multiple names for the same type. This
807 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
808 uses structural typing, the name is not part of the type. When printing out
809 LLVM IR, the printer will pick <em>one name</em> to render all types of a
810 particular shape. This means that if you have code where two different
811 source types end up having the same LLVM type, that the dumper will sometimes
812 print the "wrong" or unexpected type. This is an important design point and
813 isn't going to change.</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]814
815</div>
816
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]817<!-- ======================================================================= -->
818<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]819 <a name="globalvars">Global Variables</a>
820</div>
821
822<div class="doc_text">
823
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]824<p>Global variables define regions of memory allocated at compilation time
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]825 instead of run-time. Global variables may optionally be initialized, may
826 have an explicit section to be placed in, and may have an optional explicit
827 alignment specified. A variable may be defined as "thread_local", which
828 means that it will not be shared by threads (each thread will have a
829 separated copy of the variable). A variable may be defined as a global
830 "constant," which indicates that the contents of the variable
831 will <b>never</b> be modified (enabling better optimization, allowing the
832 global data to be placed in the read-only section of an executable, etc).
833 Note that variables that need runtime initialization cannot be marked
834 "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]835
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]836<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
837 constant, even if the final definition of the global is not. This capability
838 can be used to enable slightly better optimization of the program, but
839 requires the language definition to guarantee that optimizations based on the
840 'constantness' are valid for the translation units that do not include the
841 definition.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]842
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]843<p>As SSA values, global variables define pointer values that are in scope
844 (i.e. they dominate) all basic blocks in the program. Global variables
845 always define a pointer to their "content" type because they describe a
846 region of memory, and all memory objects in LLVM are accessed through
847 pointers.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]848
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]849<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
850 that the address is not significant, only the content. Constants marked
Rafael Espindolaa5eaa862011-01-15 08:20:57 +0000[diff] [blame]851 like this can be merged with other constants if they have the same
852 initializer. Note that a constant with significant address <em>can</em>
853 be merged with a <tt>unnamed_addr</tt> constant, the result being a
854 constant whose address is significant.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]855
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]856<p>A global variable may be declared to reside in a target-specific numbered
857 address space. For targets that support them, address spaces may affect how
858 optimizations are performed and/or what target instructions are used to
859 access the variable. The default address space is zero. The address space
860 qualifier must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000[diff] [blame]861
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]862<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]863 supports it, it will emit globals to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]864
Chris Lattnerce99fa92010-04-28 00:13:42 +0000[diff] [blame]865<p>An explicit alignment may be specified for a global, which must be a power
866 of 2. If not present, or if the alignment is set to zero, the alignment of
867 the global is set by the target to whatever it feels convenient. If an
868 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner2d4b8ee2010-04-28 00:31:12 +0000[diff] [blame]869 alignment. Targets and optimizers are not allowed to over-align the global
870 if the global has an assigned section. In this case, the extra alignment
871 could be observable: for example, code could assume that the globals are
872 densely packed in their section and try to iterate over them as an array,
873 alignment padding would break this iteration.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]874
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]875<p>For example, the following defines a global in a numbered address space with
876 an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]877
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]878<pre class="doc_code">
Dan Gohman398873c2009-01-11 00:40:00 +0000[diff] [blame]879@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]880</pre>
881
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]882</div>
883
884
885<!-- ======================================================================= -->
886<div class="doc_subsection">
887 <a name="functionstructure">Functions</a>
888</div>
889
890<div class="doc_text">
891
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]892<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]893 optional <a href="#linkage">linkage type</a>, an optional
894 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]895 <a href="#callingconv">calling convention</a>,
896 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]897 <a href="#paramattrs">parameter attribute</a> for the return type, a function
898 name, a (possibly empty) argument list (each with optional
899 <a href="#paramattrs">parameter attributes</a>), optional
900 <a href="#fnattrs">function attributes</a>, an optional section, an optional
901 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
902 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]903
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]904<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
905 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]906 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]907 <a href="#callingconv">calling convention</a>,
908 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]909 <a href="#paramattrs">parameter attribute</a> for the return type, a function
910 name, a possibly empty list of arguments, an optional alignment, and an
911 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]912
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]913<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]914 (Control Flow Graph) for the function. Each basic block may optionally start
915 with a label (giving the basic block a symbol table entry), contains a list
916 of instructions, and ends with a <a href="#terminators">terminator</a>
917 instruction (such as a branch or function return).</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]918
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]919<p>The first basic block in a function is special in two ways: it is immediately
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]920 executed on entrance to the function, and it is not allowed to have
921 predecessor basic blocks (i.e. there can not be any branches to the entry
922 block of a function). Because the block can have no predecessors, it also
923 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]924
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]925<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]926 supports it, it will emit functions to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]927
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]928<p>An explicit alignment may be specified for a function. If not present, or if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]929 the alignment is set to zero, the alignment of the function is set by the
930 target to whatever it feels convenient. If an explicit alignment is
931 specified, the function is forced to have at least that much alignment. All
932 alignments must be a power of 2.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]933
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]934<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
935 be significant and two identical functions can be merged</p>.
936
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]937<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]938<pre class="doc_code">
Chris Lattner50ad45c2008-10-13 16:55:18 +0000[diff] [blame]939define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]940 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
941 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
942 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
943 [<a href="#gc">gc</a>] { ... }
944</pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]945
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]946</div>
947
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]948<!-- ======================================================================= -->
949<div class="doc_subsection">
950 <a name="aliasstructure">Aliases</a>
951</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]952
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]953<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]954
955<p>Aliases act as "second name" for the aliasee value (which can be either
956 function, global variable, another alias or bitcast of global value). Aliases
957 may have an optional <a href="#linkage">linkage type</a>, and an
958 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]959
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]960<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]961<pre class="doc_code">
Duncan Sands0b23ac12008-09-12 20:48:21 +0000[diff] [blame]962@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]963</pre>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]964
965</div>
966
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]967<!-- ======================================================================= -->
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]968<div class="doc_subsection">
969 <a name="namedmetadatastructure">Named Metadata</a>
970</div>
971
972<div class="doc_text">
973
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]974<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]975 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]976 a named metadata.</p>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]977
978<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]979<pre class="doc_code">
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]980; Some unnamed metadata nodes, which are referenced by the named metadata.
981!0 = metadata !{metadata !"zero"}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]982!1 = metadata !{metadata !"one"}
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]983!2 = metadata !{metadata !"two"}
Dan Gohman1005bc52010-07-13 19:48:13 +0000[diff] [blame]984; A named metadata.
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]985!name = !{!0, !1, !2}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]986</pre>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]987
988</div>
989
990<!-- ======================================================================= -->
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]991<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]992
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]993<div class="doc_text">
994
995<p>The return type and each parameter of a function type may have a set of
996 <i>parameter attributes</i> associated with them. Parameter attributes are
997 used to communicate additional information about the result or parameters of
998 a function. Parameter attributes are considered to be part of the function,
999 not of the function type, so functions with different parameter attributes
1000 can have the same function type.</p>
1001
1002<p>Parameter attributes are simple keywords that follow the type specified. If
1003 multiple parameter attributes are needed, they are space separated. For
1004 example:</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1005
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1006<pre class="doc_code">
Nick Lewyckyb6a7d252009-02-15 23:06:14 +0000[diff] [blame]1007declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner66d922c2008-10-04 18:33:34 +0000[diff] [blame]1008declare i32 @atoi(i8 zeroext)
1009declare signext i8 @returns_signed_char()
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1010</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1011
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1012<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1013 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1014
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1015<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1016
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1017<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1018 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1019 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichebe81732011-03-16 22:20:18 +0000[diff] [blame]1020 should be zero-extended to the extent required by the target's ABI (which
1021 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1022 parameter) or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1023
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1024 <dt><tt><b>signext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1025 <dd>This indicates to the code generator that the parameter or return value
1026 should be sign-extended to a 32-bit value by the caller (for a parameter)
1027 or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1028
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1029 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1030 <dd>This indicates that this parameter or return value should be treated in a
1031 special target-dependent fashion during while emitting code for a function
1032 call or return (usually, by putting it in a register as opposed to memory,
1033 though some targets use it to distinguish between two different kinds of
1034 registers). Use of this attribute is target-specific.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1035
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1036 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1037 <dd><p>This indicates that the pointer parameter should really be passed by
1038 value to the function. The attribute implies that a hidden copy of the
1039 pointee
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1040 is made between the caller and the callee, so the callee is unable to
1041 modify the value in the callee. This attribute is only valid on LLVM
1042 pointer arguments. It is generally used to pass structs and arrays by
1043 value, but is also valid on pointers to scalars. The copy is considered
1044 to belong to the caller not the callee (for example,
1045 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1046 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1047 values.</p>
1048
1049 <p>The byval attribute also supports specifying an alignment with
1050 the align attribute. It indicates the alignment of the stack slot to
1051 form and the known alignment of the pointer specified to the call site. If
1052 the alignment is not specified, then the code generator makes a
1053 target-specific assumption.</p></dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1054
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1055 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1056 <dd>This indicates that the pointer parameter specifies the address of a
1057 structure that is the return value of the function in the source program.
1058 This pointer must be guaranteed by the caller to be valid: loads and
1059 stores to the structure may be assumed by the callee to not to trap. This
1060 may only be applied to the first parameter. This is not a valid attribute
1061 for return values. </dd>
1062
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1063 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1064 <dd>This indicates that pointer values
1065 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmanefca7f92010-07-02 23:46:54 +0000[diff] [blame]1066 value do not alias pointer values which are not <i>based</i> on it,
1067 ignoring certain "irrelevant" dependencies.
1068 For a call to the parent function, dependencies between memory
1069 references from before or after the call and from those during the call
1070 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1071 return value used in that call.
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1072 The caller shares the responsibility with the callee for ensuring that
1073 these requirements are met.
1074 For further details, please see the discussion of the NoAlias response in
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1075 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1076<br>
John McCall191d4ee2010-07-06 21:07:14 +0000[diff] [blame]1077 Note that this definition of <tt>noalias</tt> is intentionally
1078 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner211244a2010-07-06 20:51:35 +0000[diff] [blame]1079 arguments, though it is slightly weaker.
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1080<br>
1081 For function return values, C99's <tt>restrict</tt> is not meaningful,
1082 while LLVM's <tt>noalias</tt> is.
1083 </dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1084
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1085 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1086 <dd>This indicates that the callee does not make any copies of the pointer
1087 that outlive the callee itself. This is not a valid attribute for return
1088 values.</dd>
1089
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1090 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1091 <dd>This indicates that the pointer parameter can be excised using the
1092 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1093 attribute for return values.</dd>
1094</dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1095
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1096</div>
1097
1098<!-- ======================================================================= -->
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1099<div class="doc_subsection">
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1100 <a name="gc">Garbage Collector Names</a>
1101</div>
1102
1103<div class="doc_text">
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1104
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1105<p>Each function may specify a garbage collector name, which is simply a
1106 string:</p>
1107
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1108<pre class="doc_code">
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1109define void @f() gc "name" { ... }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1110</pre>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1111
1112<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1113 collector which will cause the compiler to alter its output in order to
1114 support the named garbage collection algorithm.</p>
1115
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1116</div>
1117
1118<!-- ======================================================================= -->
1119<div class="doc_subsection">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1120 <a name="fnattrs">Function Attributes</a>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1121</div>
1122
1123<div class="doc_text">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1124
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1125<p>Function attributes are set to communicate additional information about a
1126 function. Function attributes are considered to be part of the function, not
1127 of the function type, so functions with different parameter attributes can
1128 have the same function type.</p>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1129
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1130<p>Function attributes are simple keywords that follow the type specified. If
1131 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1132
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1133<pre class="doc_code">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1134define void @f() noinline { ... }
1135define void @f() alwaysinline { ... }
1136define void @f() alwaysinline optsize { ... }
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1137define void @f() optsize { ... }
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1138</pre>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1139
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1140<dl>
Charles Davis1e063d12010-02-12 00:31:15 +0000[diff] [blame]1141 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1142 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1143 the backend should forcibly align the stack pointer. Specify the
1144 desired alignment, which must be a power of two, in parentheses.
1145
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1146 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1147 <dd>This attribute indicates that the inliner should attempt to inline this
1148 function into callers whenever possible, ignoring any active inlining size
1149 threshold for this caller.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1150
Charles Davis970bfcc2010-10-25 15:37:09 +0000[diff] [blame]1151 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis6f12e292010-10-25 16:29:03 +0000[diff] [blame]1152 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis0076d202010-10-25 19:07:39 +0000[diff] [blame]1153 meaning the function can be patched and/or hooked even while it is
1154 loaded into memory. On x86, the function prologue will be preceded
1155 by six bytes of padding and will begin with a two-byte instruction.
1156 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1157 higher were compiled in this fashion.</dd>
Charles Davis970bfcc2010-10-25 15:37:09 +0000[diff] [blame]1158
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +0000[diff] [blame]1159 <dt><tt><b>inlinehint</b></tt></dt>
1160 <dd>This attribute indicates that the source code contained a hint that inlining
1161 this function is desirable (such as the "inline" keyword in C/C++). It
1162 is just a hint; it imposes no requirements on the inliner.</dd>
1163
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1164 <dt><tt><b>naked</b></tt></dt>
1165 <dd>This attribute disables prologue / epilogue emission for the function.
1166 This can have very system-specific consequences.</dd>
1167
1168 <dt><tt><b>noimplicitfloat</b></tt></dt>
1169 <dd>This attributes disables implicit floating point instructions.</dd>
1170
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1171 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1172 <dd>This attribute indicates that the inliner should never inline this
1173 function in any situation. This attribute may not be used together with
1174 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1175
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1176 <dt><tt><b>noredzone</b></tt></dt>
1177 <dd>This attribute indicates that the code generator should not use a red
1178 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1179
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1180 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1181 <dd>This function attribute indicates that the function never returns
1182 normally. This produces undefined behavior at runtime if the function
1183 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +0000[diff] [blame]1184
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1185 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1186 <dd>This function attribute indicates that the function never returns with an
1187 unwind or exceptional control flow. If the function does unwind, its
1188 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +0000[diff] [blame]1189
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1190 <dt><tt><b>optsize</b></tt></dt>
1191 <dd>This attribute suggests that optimization passes and code generator passes
1192 make choices that keep the code size of this function low, and otherwise
1193 do optimizations specifically to reduce code size.</dd>
1194
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1195 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1196 <dd>This attribute indicates that the function computes its result (or decides
1197 to unwind an exception) based strictly on its arguments, without
1198 dereferencing any pointer arguments or otherwise accessing any mutable
1199 state (e.g. memory, control registers, etc) visible to caller functions.
1200 It does not write through any pointer arguments
1201 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1202 changes any state visible to callers. This means that it cannot unwind
1203 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1204 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +0000[diff] [blame]1205
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1206 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1207 <dd>This attribute indicates that the function does not write through any
1208 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1209 arguments) or otherwise modify any state (e.g. memory, control registers,
1210 etc) visible to caller functions. It may dereference pointer arguments
1211 and read state that may be set in the caller. A readonly function always
1212 returns the same value (or unwinds an exception identically) when called
1213 with the same set of arguments and global state. It cannot unwind an
1214 exception by calling the <tt>C++</tt> exception throwing methods, but may
1215 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +0000[diff] [blame]1216
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1217 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1218 <dd>This attribute indicates that the function should emit a stack smashing
1219 protector. It is in the form of a "canary"&mdash;a random value placed on
1220 the stack before the local variables that's checked upon return from the
1221 function to see if it has been overwritten. A heuristic is used to
1222 determine if a function needs stack protectors or not.<br>
1223<br>
1224 If a function that has an <tt>ssp</tt> attribute is inlined into a
1225 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1226 function will have an <tt>ssp</tt> attribute.</dd>
1227
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1228 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1229 <dd>This attribute indicates that the function should <em>always</em> emit a
1230 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1231 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1232<br>
1233 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1234 function that doesn't have an <tt>sspreq</tt> attribute or which has
1235 an <tt>ssp</tt> attribute, then the resulting function will have
1236 an <tt>sspreq</tt> attribute.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1237</dl>
1238
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1239</div>
1240
1241<!-- ======================================================================= -->
1242<div class="doc_subsection">
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]1243 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1244</div>
1245
1246<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1247
1248<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1249 the GCC "file scope inline asm" blocks. These blocks are internally
1250 concatenated by LLVM and treated as a single unit, but may be separated in
1251 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1252
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1253<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1254module asm "inline asm code goes here"
1255module asm "more can go here"
1256</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1257
1258<p>The strings can contain any character by escaping non-printable characters.
1259 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1260 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1261
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1262<p>The inline asm code is simply printed to the machine code .s file when
1263 assembly code is generated.</p>
1264
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1265</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1266
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1267<!-- ======================================================================= -->
1268<div class="doc_subsection">
1269 <a name="datalayout">Data Layout</a>
1270</div>
1271
1272<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1273
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1274<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1275 data is to be laid out in memory. The syntax for the data layout is
1276 simply:</p>
1277
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1278<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1279target datalayout = "<i>layout specification</i>"
1280</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1281
1282<p>The <i>layout specification</i> consists of a list of specifications
1283 separated by the minus sign character ('-'). Each specification starts with
1284 a letter and may include other information after the letter to define some
1285 aspect of the data layout. The specifications accepted are as follows:</p>
1286
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1287<dl>
1288 <dt><tt>E</tt></dt>
1289 <dd>Specifies that the target lays out data in big-endian form. That is, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1290 bits with the most significance have the lowest address location.</dd>
1291
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1292 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1293 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1294 the bits with the least significance have the lowest address
1295 location.</dd>
1296
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1297 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1298 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1299 <i>preferred</i> alignments. All sizes are in bits. Specifying
1300 the <i>pref</i> alignment is optional. If omitted, the
1301 preceding <tt>:</tt> should be omitted too.</dd>
1302
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1303 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1304 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1305 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1306
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1307 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1308 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1309 <i>size</i>.</dd>
1310
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1311 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1312 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +0000[diff] [blame]1313 <i>size</i>. Only values of <i>size</i> that are supported by the target
1314 will work. 32 (float) and 64 (double) are supported on all targets;
1315 80 or 128 (different flavors of long double) are also supported on some
1316 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1317
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1318 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1319 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1320 <i>size</i>.</dd>
1321
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1322 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1323 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1324 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1325
1326 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1327 <dd>This specifies a set of native integer widths for the target CPU
1328 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1329 "n32:64" for PowerPC 64, or "n8:16:32:64" for X86-64. Elements of
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1330 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1331 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1332</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1333
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1334<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]1335 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1336 specifications in the <tt>datalayout</tt> keyword. The default specifications
1337 are given in this list:</p>
1338
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1339<ul>
1340 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +0000[diff] [blame]1341 <li><tt>p:64:64:64</tt> - 64-bit pointers with 64-bit alignment</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1342 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1343 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1344 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1345 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1346 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1347 alignment of 64-bits</li>
1348 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1349 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1350 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1351 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1352 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1353 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1354</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1355
1356<p>When LLVM is determining the alignment for a given type, it uses the
1357 following rules:</p>
1358
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1359<ol>
1360 <li>If the type sought is an exact match for one of the specifications, that
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1361 specification is used.</li>
1362
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1363 <li>If no match is found, and the type sought is an integer type, then the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1364 smallest integer type that is larger than the bitwidth of the sought type
1365 is used. If none of the specifications are larger than the bitwidth then
1366 the the largest integer type is used. For example, given the default
1367 specifications above, the i7 type will use the alignment of i8 (next
1368 largest) while both i65 and i256 will use the alignment of i64 (largest
1369 specified).</li>
1370
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1371 <li>If no match is found, and the type sought is a vector type, then the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1372 largest vector type that is smaller than the sought vector type will be
1373 used as a fall back. This happens because &lt;128 x double&gt; can be
1374 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1375</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1376
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1377</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1378
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1379<!-- ======================================================================= -->
1380<div class="doc_subsection">
1381 <a name="pointeraliasing">Pointer Aliasing Rules</a>
1382</div>
1383
1384<div class="doc_text">
1385
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1386<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +0000[diff] [blame]1387with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1388is undefined. Pointer values are associated with address ranges
1389according to the following rules:</p>
1390
1391<ul>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1392 <li>A pointer value is associated with the addresses associated with
1393 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1394 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1395 range of the variable's storage.</li>
1396 <li>The result value of an allocation instruction is associated with
1397 the address range of the allocated storage.</li>
1398 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1399 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1400 <li>An integer constant other than zero or a pointer value returned
1401 from a function not defined within LLVM may be associated with address
1402 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1403 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1404 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1405</ul>
1406
1407<p>A pointer value is <i>based</i> on another pointer value according
1408 to the following rules:</p>
1409
1410<ul>
1411 <li>A pointer value formed from a
1412 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1413 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1414 <li>The result value of a
1415 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1416 of the <tt>bitcast</tt>.</li>
1417 <li>A pointer value formed by an
1418 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1419 pointer values that contribute (directly or indirectly) to the
1420 computation of the pointer's value.</li>
1421 <li>The "<i>based</i> on" relationship is transitive.</li>
1422</ul>
1423
1424<p>Note that this definition of <i>"based"</i> is intentionally
1425 similar to the definition of <i>"based"</i> in C99, though it is
1426 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1427
1428<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1429<tt><a href="#i_load">load</a></tt> merely indicates the size and
1430alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +0000[diff] [blame]1431interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1432<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1433and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1434
1435<p>Consequently, type-based alias analysis, aka TBAA, aka
1436<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1437LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1438additional information which specialized optimization passes may use
1439to implement type-based alias analysis.</p>
1440
1441</div>
1442
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1443<!-- ======================================================================= -->
1444<div class="doc_subsection">
1445 <a name="volatile">Volatile Memory Accesses</a>
1446</div>
1447
1448<div class="doc_text">
1449
1450<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1451href="#i_store"><tt>store</tt></a>s, and <a
1452href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1453The optimizers must not change the number of volatile operations or change their
1454order of execution relative to other volatile operations. The optimizers
1455<i>may</i> change the order of volatile operations relative to non-volatile
1456operations. This is not Java's "volatile" and has no cross-thread
1457synchronization behavior.</p>
1458
1459</div>
1460
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1461<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1462<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1463<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1464
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1465<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1466
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1467<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1468 intermediate representation. Being typed enables a number of optimizations
1469 to be performed on the intermediate representation directly, without having
1470 to do extra analyses on the side before the transformation. A strong type
1471 system makes it easier to read the generated code and enables novel analyses
1472 and transformations that are not feasible to perform on normal three address
1473 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1474
1475</div>
1476
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1477<!-- ======================================================================= -->
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1478<div class="doc_subsection"> <a name="t_classifications">Type
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1479Classifications</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1480
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1481<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1482
1483<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1484
1485<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1486 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1487 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1488 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1489 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1490 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1491 </tr>
1492 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1493 <td><a href="#t_floating">floating point</a></td>
1494 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1495 </tr>
1496 <tr>
1497 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1498 <td><a href="#t_integer">integer</a>,
1499 <a href="#t_floating">floating point</a>,
1500 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +0000[diff] [blame]1501 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]1502 <a href="#t_struct">structure</a>,
1503 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1504 <a href="#t_label">label</a>,
1505 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1506 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1507 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1508 <tr>
1509 <td><a href="#t_primitive">primitive</a></td>
1510 <td><a href="#t_label">label</a>,
1511 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +0000[diff] [blame]1512 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1513 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1514 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1515 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1516 </tr>
1517 <tr>
1518 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1519 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1520 <a href="#t_function">function</a>,
1521 <a href="#t_pointer">pointer</a>,
1522 <a href="#t_struct">structure</a>,
1523 <a href="#t_pstruct">packed structure</a>,
1524 <a href="#t_vector">vector</a>,
1525 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +0000[diff] [blame]1526 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1527 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1528 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1529</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1530
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1531<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1532 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1533 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1534
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1535</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1536
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1537<!-- ======================================================================= -->
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1538<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1539
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1540<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1541
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1542<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1543 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1544
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1545</div>
1546
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1547<!-- _______________________________________________________________________ -->
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1548<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1549
1550<div class="doc_text">
1551
1552<h5>Overview:</h5>
1553<p>The integer type is a very simple type that simply specifies an arbitrary
1554 bit width for the integer type desired. Any bit width from 1 bit to
1555 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1556
1557<h5>Syntax:</h5>
1558<pre>
1559 iN
1560</pre>
1561
1562<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1563 value.</p>
1564
1565<h5>Examples:</h5>
1566<table class="layout">
1567 <tr class="layout">
1568 <td class="left"><tt>i1</tt></td>
1569 <td class="left">a single-bit integer.</td>
1570 </tr>
1571 <tr class="layout">
1572 <td class="left"><tt>i32</tt></td>
1573 <td class="left">a 32-bit integer.</td>
1574 </tr>
1575 <tr class="layout">
1576 <td class="left"><tt>i1942652</tt></td>
1577 <td class="left">a really big integer of over 1 million bits.</td>
1578 </tr>
1579</table>
1580
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1581</div>
1582
1583<!-- _______________________________________________________________________ -->
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1584<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1585
1586<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1587
1588<table>
1589 <tbody>
1590 <tr><th>Type</th><th>Description</th></tr>
1591 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1592 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1593 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1594 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1595 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1596 </tbody>
1597</table>
1598
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1599</div>
1600
1601<!-- _______________________________________________________________________ -->
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1602<div class="doc_subsubsection"> <a name="t_x86mmx">X86mmx Type</a> </div>
1603
1604<div class="doc_text">
1605
1606<h5>Overview:</h5>
1607<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>
1608
1609<h5>Syntax:</h5>
1610<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +0000[diff] [blame]1611 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1612</pre>
1613
1614</div>
1615
1616<!-- _______________________________________________________________________ -->
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1617<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1618
1619<div class="doc_text">
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1620
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1621<h5>Overview:</h5>
1622<p>The void type does not represent any value and has no size.</p>
1623
1624<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1625<pre>
1626 void
1627</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1628
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1629</div>
1630
1631<!-- _______________________________________________________________________ -->
1632<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1633
1634<div class="doc_text">
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1635
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1636<h5>Overview:</h5>
1637<p>The label type represents code labels.</p>
1638
1639<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1640<pre>
1641 label
1642</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1643
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1644</div>
1645
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1646<!-- _______________________________________________________________________ -->
1647<div class="doc_subsubsection"> <a name="t_metadata">Metadata Type</a> </div>
1648
1649<div class="doc_text">
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1650
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1651<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1652<p>The metadata type represents embedded metadata. No derived types may be
1653 created from metadata except for <a href="#t_function">function</a>
1654 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1655
1656<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1657<pre>
1658 metadata
1659</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1660
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1661</div>
1662
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1663
1664<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1665<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1666
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1667<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1668
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1669<p>The real power in LLVM comes from the derived types in the system. This is
1670 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1671 useful types. Each of these types contain one or more element types which
1672 may be a primitive type, or another derived type. For example, it is
1673 possible to have a two dimensional array, using an array as the element type
1674 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1675
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1676
1677</div>
1678
1679<!-- _______________________________________________________________________ -->
1680<div class="doc_subsubsection"> <a name="t_aggregate">Aggregate Types</a> </div>
1681
1682<div class="doc_text">
1683
1684<p>Aggregate Types are a subset of derived types that can contain multiple
1685 member types. <a href="#t_array">Arrays</a>,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]1686 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1687 aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1688
1689</div>
1690
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1691<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1692<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1693
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1694<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1695
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1696<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1697<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1698 sequentially in memory. The array type requires a size (number of elements)
1699 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1700
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1701<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1702<pre>
1703 [&lt;# elements&gt; x &lt;elementtype&gt;]
1704</pre>
1705
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1706<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1707 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1708
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1709<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1710<table class="layout">
1711 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1712 <td class="left"><tt>[40 x i32]</tt></td>
1713 <td class="left">Array of 40 32-bit integer values.</td>
1714 </tr>
1715 <tr class="layout">
1716 <td class="left"><tt>[41 x i32]</tt></td>
1717 <td class="left">Array of 41 32-bit integer values.</td>
1718 </tr>
1719 <tr class="layout">
1720 <td class="left"><tt>[4 x i8]</tt></td>
1721 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1722 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1723</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1724<p>Here are some examples of multidimensional arrays:</p>
1725<table class="layout">
1726 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1727 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1728 <td class="left">3x4 array of 32-bit integer values.</td>
1729 </tr>
1730 <tr class="layout">
1731 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1732 <td class="left">12x10 array of single precision floating point values.</td>
1733 </tr>
1734 <tr class="layout">
1735 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1736 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1737 </tr>
1738</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1739
Dan Gohman7657f6b2009-11-09 19:01:53 +0000[diff] [blame]1740<p>There is no restriction on indexing beyond the end of the array implied by
1741 a static type (though there are restrictions on indexing beyond the bounds
1742 of an allocated object in some cases). This means that single-dimension
1743 'variable sized array' addressing can be implemented in LLVM with a zero
1744 length array type. An implementation of 'pascal style arrays' in LLVM could
1745 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1746
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1747</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1748
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1749<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1750<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1751
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1752<div class="doc_text">
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1753
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1754<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1755<p>The function type can be thought of as a function signature. It consists of
1756 a return type and a list of formal parameter types. The return type of a
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]1757 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]1758
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1759<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1760<pre>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]1761 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1762</pre>
1763
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1764<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1765 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1766 which indicates that the function takes a variable number of arguments.
1767 Variable argument functions can access their arguments with
1768 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]1769 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1770 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1771
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1772<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1773<table class="layout">
1774 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1775 <td class="left"><tt>i32 (i32)</tt></td>
1776 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1777 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1778 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]1779 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]1780 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1781 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]1782 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1783 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1784 </td>
1785 </tr><tr class="layout">
1786 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1787 <td class="left">A vararg function that takes at least one
1788 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1789 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1790 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1791 </td>
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]1792 </tr><tr class="layout">
1793 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]1794 <td class="left">A function taking an <tt>i32</tt>, returning a
1795 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]1796 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1797 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1798</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1799
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1800</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1801
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1802<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1803<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1804
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1805<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1806
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1807<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1808<p>The structure type is used to represent a collection of data members together
1809 in memory. The packing of the field types is defined to match the ABI of the
1810 underlying processor. The elements of a structure may be any type that has a
1811 size.</p>
1812
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]1813<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1814 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1815 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1816 Structures in registers are accessed using the
1817 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1818 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1819<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1820<pre>
1821 { &lt;type list&gt; }
1822</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1823
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1824<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1825<table class="layout">
1826 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1827 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1828 <td class="left">A triple of three <tt>i32</tt> values</td>
1829 </tr><tr class="layout">
1830 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1831 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1832 second element is a <a href="#t_pointer">pointer</a> to a
1833 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1834 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1835 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1836</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1837
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1838</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1839
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1840<!-- _______________________________________________________________________ -->
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1841<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1842</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1843
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1844<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1845
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1846<h5>Overview:</h5>
1847<p>The packed structure type is used to represent a collection of data members
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1848 together in memory. There is no padding between fields. Further, the
1849 alignment of a packed structure is 1 byte. The elements of a packed
1850 structure may be any type that has a size.</p>
1851
1852<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt> and
1853 '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field with
1854 the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
1855
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1856<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1857<pre>
1858 &lt; { &lt;type list&gt; } &gt;
1859</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1860
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1861<h5>Examples:</h5>
1862<table class="layout">
1863 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1864 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1865 <td class="left">A triple of three <tt>i32</tt> values</td>
1866 </tr><tr class="layout">
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1867 <td class="left">
1868<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1869 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1870 second element is a <a href="#t_pointer">pointer</a> to a
1871 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1872 an <tt>i32</tt>.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1873 </tr>
1874</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1875
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1876</div>
1877
1878<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1879<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]1880
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1881<div class="doc_text">
1882
1883<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +0000[diff] [blame]1884<p>The pointer type is used to specify memory locations.
1885 Pointers are commonly used to reference objects in memory.</p>
1886
1887<p>Pointer types may have an optional address space attribute defining the
1888 numbered address space where the pointed-to object resides. The default
1889 address space is number zero. The semantics of non-zero address
1890 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1891
1892<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
1893 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]1894
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1895<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1896<pre>
1897 &lt;type&gt; *
1898</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1899
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1900<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1901<table class="layout">
1902 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]1903 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1904 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1905 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1906 </tr>
1907 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]1908 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1909 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1910 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1911 <tt>i32</tt>.</td>
1912 </tr>
1913 <tr class="layout">
1914 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1915 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1916 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1917 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1918</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1919
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1920</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1921
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1922<!-- _______________________________________________________________________ -->
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1923<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1924
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1925<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1926
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1927<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1928<p>A vector type is a simple derived type that represents a vector of elements.
1929 Vector types are used when multiple primitive data are operated in parallel
1930 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +0000[diff] [blame]1931 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1932 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1933
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1934<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1935<pre>
1936 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1937</pre>
1938
Chris Lattner7d2e7be2010-10-10 18:20:35 +0000[diff] [blame]1939<p>The number of elements is a constant integer value larger than 0; elementtype
1940 may be any integer or floating point type. Vectors of size zero are not
1941 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1942
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1943<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1944<table class="layout">
1945 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1946 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1947 <td class="left">Vector of 4 32-bit integer values.</td>
1948 </tr>
1949 <tr class="layout">
1950 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1951 <td class="left">Vector of 8 32-bit floating-point values.</td>
1952 </tr>
1953 <tr class="layout">
1954 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1955 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1956 </tr>
1957</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1958
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1959</div>
1960
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1961<!-- _______________________________________________________________________ -->
1962<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1963<div class="doc_text">
1964
1965<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1966<p>Opaque types are used to represent unknown types in the system. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1967 corresponds (for example) to the C notion of a forward declared structure
1968 type. In LLVM, opaque types can eventually be resolved to any type (not just
1969 a structure type).</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1970
1971<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1972<pre>
1973 opaque
1974</pre>
1975
1976<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1977<table class="layout">
1978 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1979 <td class="left"><tt>opaque</tt></td>
1980 <td class="left">An opaque type.</td>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1981 </tr>
1982</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1983
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1984</div>
1985
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]1986<!-- ======================================================================= -->
1987<div class="doc_subsection">
1988 <a name="t_uprefs">Type Up-references</a>
1989</div>
1990
1991<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1992
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]1993<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1994<p>An "up reference" allows you to refer to a lexically enclosing type without
1995 requiring it to have a name. For instance, a structure declaration may
1996 contain a pointer to any of the types it is lexically a member of. Example
1997 of up references (with their equivalent as named type declarations)
1998 include:</p>
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]1999
2000<pre>
Chris Lattner3060f5b2009-02-09 10:00:56 +0000[diff] [blame]2001 { \2 * } %x = type { %x* }
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]2002 { \2 }* %y = type { %y }*
2003 \1* %z = type %z*
2004</pre>
2005
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2006<p>An up reference is needed by the asmprinter for printing out cyclic types
2007 when there is no declared name for a type in the cycle. Because the
2008 asmprinter does not want to print out an infinite type string, it needs a
2009 syntax to handle recursive types that have no names (all names are optional
2010 in llvm IR).</p>
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]2011
2012<h5>Syntax:</h5>
2013<pre>
2014 \&lt;level&gt;
2015</pre>
2016
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2017<p>The level is the count of the lexical type that is being referred to.</p>
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]2018
2019<h5>Examples:</h5>
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]2020<table class="layout">
2021 <tr class="layout">
2022 <td class="left"><tt>\1*</tt></td>
2023 <td class="left">Self-referential pointer.</td>
2024 </tr>
2025 <tr class="layout">
2026 <td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
2027 <td class="left">Recursive structure where the upref refers to the out-most
2028 structure.</td>
2029 </tr>
2030</table>
Chris Lattner242d61d2009-02-02 07:32:36 +0000[diff] [blame]2031
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2032</div>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2033
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2034<!-- *********************************************************************** -->
2035<div class="doc_section"> <a name="constants">Constants</a> </div>
2036<!-- *********************************************************************** -->
2037
2038<div class="doc_text">
2039
2040<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2041 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2042
2043</div>
2044
2045<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]2046<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2047
2048<div class="doc_text">
2049
2050<dl>
2051 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2052 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]2053 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2054
2055 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2056 <dd>Standard integers (such as '4') are constants of
2057 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2058 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2059
2060 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2061 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2062 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2063 notation (see below). The assembler requires the exact decimal value of a
2064 floating-point constant. For example, the assembler accepts 1.25 but
2065 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2066 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2067
2068 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2069 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2070 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2071</dl>
2072
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2073<p>The one non-intuitive notation for constants is the hexadecimal form of
2074 floating point constants. For example, the form '<tt>double
2075 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2076 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2077 constants are required (and the only time that they are generated by the
2078 disassembler) is when a floating point constant must be emitted but it cannot
2079 be represented as a decimal floating point number in a reasonable number of
2080 digits. For example, NaN's, infinities, and other special values are
2081 represented in their IEEE hexadecimal format so that assembly and disassembly
2082 do not cause any bits to change in the constants.</p>
2083
Dale Johannesenbd5e5a82009-02-11 22:14:51 +0000[diff] [blame]2084<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2085 represented using the 16-digit form shown above (which matches the IEEE754
2086 representation for double); float values must, however, be exactly
2087 representable as IEE754 single precision. Hexadecimal format is always used
2088 for long double, and there are three forms of long double. The 80-bit format
2089 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2090 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2091 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2092 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2093 currently supported target uses this format. Long doubles will only work if
2094 they match the long double format on your target. All hexadecimal formats
2095 are big-endian (sign bit at the left).</p>
2096
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]2097<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2098</div>
2099
2100<!-- ======================================================================= -->
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2101<div class="doc_subsection">
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]2102<a name="aggregateconstants"></a> <!-- old anchor -->
2103<a name="complexconstants">Complex Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2104</div>
2105
2106<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2107
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2108<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2109 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2110
2111<dl>
2112 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2113 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2114 type definitions (a comma separated list of elements, surrounded by braces
2115 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2116 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2117 Structure constants must have <a href="#t_struct">structure type</a>, and
2118 the number and types of elements must match those specified by the
2119 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2120
2121 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2122 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2123 definitions (a comma separated list of elements, surrounded by square
2124 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2125 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2126 the number and types of elements must match those specified by the
2127 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2128
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2129 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2130 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2131 definitions (a comma separated list of elements, surrounded by
2132 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2133 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2134 have <a href="#t_vector">vector type</a>, and the number and types of
2135 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2136
2137 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2138 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]2139 value to zero of <em>any</em> type, including scalar and
2140 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2141 This is often used to avoid having to print large zero initializers
2142 (e.g. for large arrays) and is always exactly equivalent to using explicit
2143 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +0000[diff] [blame]2144
2145 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +0000[diff] [blame]2146 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2147 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2148 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2149 be interpreted as part of the instruction stream, metadata is a place to
2150 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2151</dl>
2152
2153</div>
2154
2155<!-- ======================================================================= -->
2156<div class="doc_subsection">
2157 <a name="globalconstants">Global Variable and Function Addresses</a>
2158</div>
2159
2160<div class="doc_text">
2161
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2162<p>The addresses of <a href="#globalvars">global variables</a>
2163 and <a href="#functionstructure">functions</a> are always implicitly valid
2164 (link-time) constants. These constants are explicitly referenced when
2165 the <a href="#identifiers">identifier for the global</a> is used and always
2166 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2167 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2168
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2169<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]2170@X = global i32 17
2171@Y = global i32 42
2172@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2173</pre>
2174
2175</div>
2176
2177<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]2178<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2179<div class="doc_text">
2180
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2181<p>The string '<tt>undef</tt>' can be used anywhere a constant is expected, and
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]2182 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2183 Undefined values may be of any type (other than '<tt>label</tt>'
2184 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2185
Chris Lattnerc608cb12009-09-11 01:49:31 +0000[diff] [blame]2186<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2187 program is well defined no matter what value is used. This gives the
2188 compiler more freedom to optimize. Here are some examples of (potentially
2189 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2190
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2191
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2192<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2193 %A = add %X, undef
2194 %B = sub %X, undef
2195 %C = xor %X, undef
2196Safe:
2197 %A = undef
2198 %B = undef
2199 %C = undef
2200</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2201
2202<p>This is safe because all of the output bits are affected by the undef bits.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2203 Any output bit can have a zero or one depending on the input bits.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2204
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2205<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2206 %A = or %X, undef
2207 %B = and %X, undef
2208Safe:
2209 %A = -1
2210 %B = 0
2211Unsafe:
2212 %A = undef
2213 %B = undef
2214</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2215
2216<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2217 For example, if <tt>%X</tt> has a zero bit, then the output of the
2218 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2219 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2220 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2221 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2222 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2223 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2224 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2225
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2226<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2227 %A = select undef, %X, %Y
2228 %B = select undef, 42, %Y
2229 %C = select %X, %Y, undef
2230Safe:
2231 %A = %X (or %Y)
2232 %B = 42 (or %Y)
2233 %C = %Y
2234Unsafe:
2235 %A = undef
2236 %B = undef
2237 %C = undef
2238</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2239
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2240<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2241 branch) conditions can go <em>either way</em>, but they have to come from one
2242 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2243 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2244 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2245 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2246 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2247 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2248
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2249<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2250 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2251
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2252 %B = undef
2253 %C = xor %B, %B
2254
2255 %D = undef
2256 %E = icmp lt %D, 4
2257 %F = icmp gte %D, 4
2258
2259Safe:
2260 %A = undef
2261 %B = undef
2262 %C = undef
2263 %D = undef
2264 %E = undef
2265 %F = undef
2266</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2267
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2268<p>This example points out that two '<tt>undef</tt>' operands are not
2269 necessarily the same. This can be surprising to people (and also matches C
2270 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2271 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2272 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2273 its value over its "live range". This is true because the variable doesn't
2274 actually <em>have a live range</em>. Instead, the value is logically read
2275 from arbitrary registers that happen to be around when needed, so the value
2276 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2277 need to have the same semantics or the core LLVM "replace all uses with"
2278 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2279
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2280<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2281 %A = fdiv undef, %X
2282 %B = fdiv %X, undef
2283Safe:
2284 %A = undef
2285b: unreachable
2286</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2287
2288<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2289 value</em> and <em>undefined behavior</em>. An undefined value (like
2290 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2291 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2292 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2293 defined on SNaN's. However, in the second example, we can make a more
2294 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2295 arbitrary value, we are allowed to assume that it could be zero. Since a
2296 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2297 the operation does not execute at all. This allows us to delete the divide and
2298 all code after it. Because the undefined operation "can't happen", the
2299 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2300
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2301<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2302a: store undef -> %X
2303b: store %X -> undef
2304Safe:
2305a: &lt;deleted&gt;
2306b: unreachable
2307</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2308
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2309<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2310 undefined value can be assumed to not have any effect; we can assume that the
2311 value is overwritten with bits that happen to match what was already there.
2312 However, a store <em>to</em> an undefined location could clobber arbitrary
2313 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2314
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2315</div>
2316
2317<!-- ======================================================================= -->
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2318<div class="doc_subsection"><a name="trapvalues">Trap Values</a></div>
2319<div class="doc_text">
2320
Dan Gohmanc68ce062010-04-26 20:21:21 +0000[diff] [blame]2321<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2322 instead of representing an unspecified bit pattern, they represent the
2323 fact that an instruction or constant expression which cannot evoke side
2324 effects has nevertheless detected a condition which results in undefined
Dan Gohmanc68ce062010-04-26 20:21:21 +0000[diff] [blame]2325 behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2326
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2327<p>There is currently no way of representing a trap value in the IR; they
Dan Gohman855abed2010-05-03 14:51:43 +0000[diff] [blame]2328 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2329 <a href="#i_add"><tt>add</tt></a> with the <tt>nsw</tt> flag.</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2330
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2331<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2332
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2333<ul>
2334<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2335 their operands.</li>
2336
2337<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2338 to their dynamic predecessor basic block.</li>
2339
2340<li>Function arguments depend on the corresponding actual argument values in
2341 the dynamic callers of their functions.</li>
2342
2343<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2344 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2345 control back to them.</li>
2346
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2347<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2348 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2349 or exception-throwing call instructions that dynamically transfer control
2350 back to them.</li>
2351
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2352<li>Non-volatile loads and stores depend on the most recent stores to all of the
2353 referenced memory addresses, following the order in the IR
2354 (including loads and stores implied by intrinsics such as
2355 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2356
Dan Gohman7c24ff12010-05-03 14:59:34 +0000[diff] [blame]2357<!-- TODO: In the case of multiple threads, this only applies if the store
2358 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2359
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2360<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2361
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2362<li>An instruction with externally visible side effects depends on the most
2363 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]2364 the order in the IR. (This includes
2365 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2366
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2367<li>An instruction <i>control-depends</i> on a
2368 <a href="#terminators">terminator instruction</a>
2369 if the terminator instruction has multiple successors and the instruction
2370 is always executed when control transfers to one of the successors, and
2371 may not be executed when control is transfered to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2372
2373<li>Dependence is transitive.</li>
2374
2375</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2376
2377<p>Whenever a trap value is generated, all values which depend on it evaluate
2378 to trap. If they have side effects, the evoke their side effects as if each
2379 operand with a trap value were undef. If they have externally-visible side
2380 effects, the behavior is undefined.</p>
2381
2382<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +0000[diff] [blame]2383
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2384<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2385entry:
2386 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2387 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2388 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2389 store i32 0, i32* %trap_yet_again ; undefined behavior
2390
2391 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2392 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2393
2394 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2395
2396 %narrowaddr = bitcast i32* @g to i16*
2397 %wideaddr = bitcast i32* @g to i64*
2398 %trap3 = load 16* %narrowaddr ; Returns a trap value.
2399 %trap4 = load i64* %widaddr ; Returns a trap value.
2400
2401 %cmp = icmp i32 slt %trap, 0 ; Returns a trap value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2402 %br i1 %cmp, %true, %end ; Branch to either destination.
2403
2404true:
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2405 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2406 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2407 br label %end
2408
2409end:
2410 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2411 ; Both edges into this PHI are
2412 ; control-dependent on %cmp, so this
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2413 ; always results in a trap value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2414
2415 volatile store i32 0, i32* @g ; %end is control-equivalent to %entry
2416 ; so this is defined (ignoring earlier
2417 ; undefined behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2418</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2419
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2420</div>
2421
2422<!-- ======================================================================= -->
Chris Lattnerf9d078e2009-10-27 21:19:13 +0000[diff] [blame]2423<div class="doc_subsection"><a name="blockaddress">Addresses of Basic
2424 Blocks</a></div>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2425<div class="doc_text">
2426
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2427<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2428
2429<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +0000[diff] [blame]2430 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2431 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2432
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2433<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2434 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2435 comparisons against null. Pointer equality tests between labels addresses
2436 results in undefined behavior &mdash; though, again, comparison against null
2437 is ok, and no label is equal to the null pointer. This may be passed around
2438 as an opaque pointer sized value as long as the bits are not inspected. This
2439 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2440 long as the original value is reconstituted before the <tt>indirectbr</tt>
2441 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2442
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2443<p>Finally, some targets may provide defined semantics when using the value as
2444 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2445
2446</div>
2447
2448
2449<!-- ======================================================================= -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2450<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
2451</div>
2452
2453<div class="doc_text">
2454
2455<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2456 to be used as constants. Constant expressions may be of
2457 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2458 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2459 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2460
2461<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2462 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2463 <dd>Truncate a constant to another type. The bit size of CST must be larger
2464 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2465
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2466 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2467 <dd>Zero extend a constant to another type. The bit size of CST must be
Duncan Sands28afd432010-07-13 12:06:14 +0000[diff] [blame]2468 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2469
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2470 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2471 <dd>Sign extend a constant to another type. The bit size of CST must be
Duncan Sands28afd432010-07-13 12:06:14 +0000[diff] [blame]2472 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2473
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2474 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2475 <dd>Truncate a floating point constant to another floating point type. The
2476 size of CST must be larger than the size of TYPE. Both types must be
2477 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2478
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2479 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2480 <dd>Floating point extend a constant to another type. The size of CST must be
2481 smaller or equal to the size of TYPE. Both types must be floating
2482 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2483
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2484 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2485 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2486 constant. TYPE must be a scalar or vector integer type. CST must be of
2487 scalar or vector floating point type. Both CST and TYPE must be scalars,
2488 or vectors of the same number of elements. If the value won't fit in the
2489 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2490
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2491 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2492 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2493 constant. TYPE must be a scalar or vector integer type. CST must be of
2494 scalar or vector floating point type. Both CST and TYPE must be scalars,
2495 or vectors of the same number of elements. If the value won't fit in the
2496 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2497
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2498 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2499 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2500 constant. TYPE must be a scalar or vector floating point type. CST must be
2501 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2502 vectors of the same number of elements. If the value won't fit in the
2503 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2504
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2505 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2506 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2507 constant. TYPE must be a scalar or vector floating point type. CST must be
2508 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2509 vectors of the same number of elements. If the value won't fit in the
2510 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2511
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2512 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2513 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2514 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2515 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2516 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2517
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2518 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2519 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2520 type. CST must be of integer type. The CST value is zero extended,
2521 truncated, or unchanged to make it fit in a pointer size. This one is
2522 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2523
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2524 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +0000[diff] [blame]2525 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2526 are the same as those for the <a href="#i_bitcast">bitcast
2527 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2528
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2529 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2530 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2531 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2532 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2533 instruction, the index list may have zero or more indexes, which are
2534 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2535
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2536 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2537 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2538
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2539 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2540 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2541
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2542 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2543 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2544
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2545 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2546 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2547 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2548
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2549 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2550 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2551 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2552
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2553 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2554 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2555 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2556
Nick Lewycky9e130ce2010-05-29 06:44:15 +0000[diff] [blame]2557 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2558 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2559 constants. The index list is interpreted in a similar manner as indices in
2560 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2561 index value must be specified.</dd>
2562
2563 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2564 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2565 constants. The index list is interpreted in a similar manner as indices in
2566 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2567 index value must be specified.</dd>
2568
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2569 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2570 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2571 be any of the <a href="#binaryops">binary</a>
2572 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2573 on operands are the same as those for the corresponding instruction
2574 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2575</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2576
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2577</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]2578
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2579<!-- *********************************************************************** -->
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2580<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
2581<!-- *********************************************************************** -->
2582
2583<!-- ======================================================================= -->
2584<div class="doc_subsection">
2585<a name="inlineasm">Inline Assembler Expressions</a>
2586</div>
2587
2588<div class="doc_text">
2589
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2590<p>LLVM supports inline assembler expressions (as opposed
2591 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2592 a special value. This value represents the inline assembler as a string
2593 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2594 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2595 expression has side effects, and a flag indicating whether the function
2596 containing the asm needs to align its stack conservatively. An example
2597 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2598
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2599<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2600i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2601</pre>
2602
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2603<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2604 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2605 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2606
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2607<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2608%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2609</pre>
2610
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2611<p>Inline asms with side effects not visible in the constraint list must be
2612 marked as having side effects. This is done through the use of the
2613 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2614
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2615<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2616call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2617</pre>
2618
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2619<p>In some cases inline asms will contain code that will not work unless the
2620 stack is aligned in some way, such as calls or SSE instructions on x86,
2621 yet will not contain code that does that alignment within the asm.
2622 The compiler should make conservative assumptions about what the asm might
2623 contain and should generate its usual stack alignment code in the prologue
2624 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2625
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2626<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2627call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2628</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2629
2630<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2631 first.</p>
2632
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2633<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2634 documented here. Constraints on what can be done (e.g. duplication, moving,
2635 etc need to be documented). This is probably best done by reference to
2636 another document that covers inline asm from a holistic perspective.</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2637</div>
2638
2639<div class="doc_subsubsection">
2640<a name="inlineasm_md">Inline Asm Metadata</a>
2641</div>
2642
2643<div class="doc_text">
2644
2645<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2646 attached to it that contains a list of constant integers. If present, the
2647 code generator will use the integer as the location cookie value when report
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2648 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]2649 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2650 source code that produced it. For example:</p>
2651
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2652<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2653call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2654...
2655!42 = !{ i32 1234567 }
2656</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2657
2658<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 +0000[diff] [blame]2659 IR. If the MDNode contains multiple constants, the code generator will use
2660 the one that corresponds to the line of the asm that the error occurs on.</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2661
2662</div>
2663
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2664<!-- ======================================================================= -->
2665<div class="doc_subsection"><a name="metadata">Metadata Nodes and Metadata
2666 Strings</a>
2667</div>
2668
2669<div class="doc_text">
2670
2671<p>LLVM IR allows metadata to be attached to instructions in the program that
2672 can convey extra information about the code to the optimizers and code
2673 generator. One example application of metadata is source-level debug
2674 information. There are two metadata primitives: strings and nodes. All
2675 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2676 preceding exclamation point ('<tt>!</tt>').</p>
2677
2678<p>A metadata string is a string surrounded by double quotes. It can contain
2679 any character by escaping non-printable characters with "\xx" where "xx" is
2680 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2681
2682<p>Metadata nodes are represented with notation similar to structure constants
2683 (a comma separated list of elements, surrounded by braces and preceded by an
2684 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2685 10}</tt>". Metadata nodes can have any values as their operand.</p>
2686
2687<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2688 metadata nodes, which can be looked up in the module symbol table. For
2689 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2690
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2691<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2692 function is using two metadata arguments.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2693
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2694<div class="doc_code">
2695<pre>
2696call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2697</pre>
2698</div>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2699
2700<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2701 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2702
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2703<div class="doc_code">
2704<pre>
2705%indvar.next = add i64 %indvar, 1, !dbg !21
2706</pre>
2707</div>
2708
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2709</div>
2710
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2711
2712<!-- *********************************************************************** -->
2713<div class="doc_section">
2714 <a name="intrinsic_globals">Intrinsic Global Variables</a>
2715</div>
2716<!-- *********************************************************************** -->
2717
2718<p>LLVM has a number of "magic" global variables that contain data that affect
2719code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2720of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2721section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2722by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2723
2724<!-- ======================================================================= -->
2725<div class="doc_subsection">
2726<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
2727</div>
2728
2729<div class="doc_text">
2730
2731<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2732href="#linkage_appending">appending linkage</a>. This array contains a list of
2733pointers to global variables and functions which may optionally have a pointer
2734cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2735
2736<pre>
2737 @X = global i8 4
2738 @Y = global i32 123
2739
2740 @llvm.used = appending global [2 x i8*] [
2741 i8* @X,
2742 i8* bitcast (i32* @Y to i8*)
2743 ], section "llvm.metadata"
2744</pre>
2745
2746<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2747compiler, assembler, and linker are required to treat the symbol as if there is
2748a reference to the global that it cannot see. For example, if a variable has
2749internal linkage and no references other than that from the <tt>@llvm.used</tt>
2750list, it cannot be deleted. This is commonly used to represent references from
2751inline asms and other things the compiler cannot "see", and corresponds to
2752"attribute((used))" in GNU C.</p>
2753
2754<p>On some targets, the code generator must emit a directive to the assembler or
2755object file to prevent the assembler and linker from molesting the symbol.</p>
2756
2757</div>
2758
2759<!-- ======================================================================= -->
2760<div class="doc_subsection">
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2761<a name="intg_compiler_used">The '<tt>llvm.compiler.used</tt>' Global Variable</a>
2762</div>
2763
2764<div class="doc_text">
2765
2766<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2767<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2768touching the symbol. On targets that support it, this allows an intelligent
2769linker to optimize references to the symbol without being impeded as it would be
2770by <tt>@llvm.used</tt>.</p>
2771
2772<p>This is a rare construct that should only be used in rare circumstances, and
2773should not be exposed to source languages.</p>
2774
2775</div>
2776
2777<!-- ======================================================================= -->
2778<div class="doc_subsection">
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2779<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
2780</div>
2781
2782<div class="doc_text">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2783<pre>
2784%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]2785@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2786</pre>
2787<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.
2788</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2789
2790</div>
2791
2792<!-- ======================================================================= -->
2793<div class="doc_subsection">
2794<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
2795</div>
2796
2797<div class="doc_text">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2798<pre>
2799%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]2800@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2801</pre>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2802
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2803<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.
2804</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2805
2806</div>
2807
2808
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2809<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2810<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
2811<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2812
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2813<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2814
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2815<p>The LLVM instruction set consists of several different classifications of
2816 instructions: <a href="#terminators">terminator
2817 instructions</a>, <a href="#binaryops">binary instructions</a>,
2818 <a href="#bitwiseops">bitwise binary instructions</a>,
2819 <a href="#memoryops">memory instructions</a>, and
2820 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2821
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2822</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2823
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2824<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2825<div class="doc_subsection"> <a name="terminators">Terminator
2826Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2827
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2828<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2829
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2830<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2831 in a program ends with a "Terminator" instruction, which indicates which
2832 block should be executed after the current block is finished. These
2833 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2834 control flow, not values (the one exception being the
2835 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2836
Duncan Sands83821c82010-04-15 20:35:54 +0000[diff] [blame]2837<p>There are seven different terminator instructions: the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2838 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2839 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2840 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling21c346e2009-11-02 00:25:26 +0000[diff] [blame]2841 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2842 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2843 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2844 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2845
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2846</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2847
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2848<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2849<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
2850Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2851
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2852<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2853
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2854<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]2855<pre>
2856 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2857 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2858</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2859
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2860<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2861<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2862 a value) from a function back to the caller.</p>
2863
2864<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2865 value and then causes control flow, and one that just causes control flow to
2866 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2867
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2868<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2869<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2870 return value. The type of the return value must be a
2871 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]2872
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2873<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2874 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2875 value or a return value with a type that does not match its type, or if it
2876 has a void return type and contains a '<tt>ret</tt>' instruction with a
2877 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2878
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2879<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2880<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2881 the calling function's context. If the caller is a
2882 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2883 instruction after the call. If the caller was an
2884 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2885 the beginning of the "normal" destination block. If the instruction returns
2886 a value, that value shall set the call or invoke instruction's return
2887 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2888
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2889<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2890<pre>
2891 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2892 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +0000[diff] [blame]2893 ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2894</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]2895
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2896</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2897<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2898<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2899
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2900<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2901
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2902<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2903<pre>
2904 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 +0000[diff] [blame]2905</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2906
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2907<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2908<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
2909 different basic block in the current function. There are two forms of this
2910 instruction, corresponding to a conditional branch and an unconditional
2911 branch.</p>
2912
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2913<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2914<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
2915 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
2916 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
2917 target.</p>
2918
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2919<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]2920<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2921 argument is evaluated. If the value is <tt>true</tt>, control flows to the
2922 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2923 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
2924
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2925<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2926<pre>
2927Test:
2928 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
2929 br i1 %cond, label %IfEqual, label %IfUnequal
2930IfEqual:
2931 <a href="#i_ret">ret</a> i32 1
2932IfUnequal:
2933 <a href="#i_ret">ret</a> i32 0
2934</pre>
2935
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2936</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2937
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2938<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2939<div class="doc_subsubsection">
2940 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2941</div>
2942
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2943<div class="doc_text">
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2944
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2945<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2946<pre>
2947 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2948</pre>
2949
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2950<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2951<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2952 several different places. It is a generalization of the '<tt>br</tt>'
2953 instruction, allowing a branch to occur to one of many possible
2954 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2955
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2956<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2957<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2958 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
2959 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
2960 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2961
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2962<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2963<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2964 destinations. When the '<tt>switch</tt>' instruction is executed, this table
2965 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]2966 transferred to the corresponding destination; otherwise, control flow is
2967 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2968
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2969<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2970<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2971 <tt>switch</tt> instruction, this instruction may be code generated in
2972 different ways. For example, it could be generated as a series of chained
2973 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2974
2975<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2976<pre>
2977 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]2978 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2979 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2980
2981 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2982 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]2983
2984 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2985 switch i32 %val, label %otherwise [ i32 0, label %onzero
2986 i32 1, label %onone
2987 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2988</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2989
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2990</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]2991
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]2992
2993<!-- _______________________________________________________________________ -->
2994<div class="doc_subsubsection">
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]2995 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]2996</div>
2997
2998<div class="doc_text">
2999
3000<h5>Syntax:</h5>
3001<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3002 indirectbr &lt;somety&gt;* &lt;address&gt;, [ label &lt;dest1&gt;, label &lt;dest2&gt;, ... ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3003</pre>
3004
3005<h5>Overview:</h5>
3006
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3007<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3008 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]3009 "<tt>address</tt>". Address must be derived from a <a
3010 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3011
3012<h5>Arguments:</h5>
3013
3014<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3015 rest of the arguments indicate the full set of possible destinations that the
3016 address may point to. Blocks are allowed to occur multiple times in the
3017 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3018
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3019<p>This destination list is required so that dataflow analysis has an accurate
3020 understanding of the CFG.</p>
3021
3022<h5>Semantics:</h5>
3023
3024<p>Control transfers to the block specified in the address argument. All
3025 possible destination blocks must be listed in the label list, otherwise this
3026 instruction has undefined behavior. This implies that jumps to labels
3027 defined in other functions have undefined behavior as well.</p>
3028
3029<h5>Implementation:</h5>
3030
3031<p>This is typically implemented with a jump through a register.</p>
3032
3033<h5>Example:</h5>
3034<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3035 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3036</pre>
3037
3038</div>
3039
3040
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3041<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3042<div class="doc_subsubsection">
3043 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
3044</div>
3045
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3046<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3047
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3048<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3049<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3050 &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 +0000[diff] [blame]3051 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3052</pre>
3053
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3054<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3055<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3056 function, with the possibility of control flow transfer to either the
3057 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3058 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3059 control flow will return to the "normal" label. If the callee (or any
3060 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3061 instruction, control is interrupted and continued at the dynamically nearest
3062 "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3063
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3064<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3065<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3066
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3067<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3068 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3069 convention</a> the call should use. If none is specified, the call
3070 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3071
3072 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3073 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3074 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3075
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3076 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3077 function value being invoked. In most cases, this is a direct function
3078 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3079 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3080
3081 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3082 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3083
3084 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]3085 signature argument types and parameter attributes. All arguments must be
3086 of <a href="#t_firstclass">first class</a> type. If the function
3087 signature indicates the function accepts a variable number of arguments,
3088 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3089
3090 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3091 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3092
3093 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3094 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3095
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3096 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3097 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3098 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3099</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3100
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3101<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3102<p>This instruction is designed to operate as a standard
3103 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3104 primary difference is that it establishes an association with a label, which
3105 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3106
3107<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3108 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3109 exception. Additionally, this is important for implementation of
3110 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3111
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3112<p>For the purposes of the SSA form, the definition of the value returned by the
3113 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3114 block to the "normal" label. If the callee unwinds then no return value is
3115 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +0000[diff] [blame]3116
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3117<p>Note that the code generator does not yet completely support unwind, and
3118that the invoke/unwind semantics are likely to change in future versions.</p>
3119
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3120<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3121<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3122 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3123 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3124 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3125 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3126</pre>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3127
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3128</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3129
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3130<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3131
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3132<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
3133Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3134
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3135<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3136
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3137<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3138<pre>
3139 unwind
3140</pre>
3141
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3142<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3143<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3144 at the first callee in the dynamic call stack which used
3145 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3146 This is primarily used to implement exception handling.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3147
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3148<h5>Semantics:</h5>
Chris Lattner72ed2002008-04-19 21:01:16 +0000[diff] [blame]3149<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3150 immediately halt. The dynamic call stack is then searched for the
3151 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3152 Once found, execution continues at the "exceptional" destination block
3153 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3154 instruction in the dynamic call chain, undefined behavior results.</p>
3155
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3156<p>Note that the code generator does not yet completely support unwind, and
3157that the invoke/unwind semantics are likely to change in future versions.</p>
3158
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3159</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3160
3161<!-- _______________________________________________________________________ -->
3162
3163<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
3164Instruction</a> </div>
3165
3166<div class="doc_text">
3167
3168<h5>Syntax:</h5>
3169<pre>
3170 unreachable
3171</pre>
3172
3173<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3174<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3175 instruction is used to inform the optimizer that a particular portion of the
3176 code is not reachable. This can be used to indicate that the code after a
3177 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3178
3179<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3180<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3181
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3182</div>
3183
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3184<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3185<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3186
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3187<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3188
3189<p>Binary operators are used to do most of the computation in a program. They
3190 require two operands of the same type, execute an operation on them, and
3191 produce a single value. The operands might represent multiple data, as is
3192 the case with the <a href="#t_vector">vector</a> data type. The result value
3193 has the same type as its operands.</p>
3194
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3195<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3196
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3197</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3198
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3199<!-- _______________________________________________________________________ -->
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3200<div class="doc_subsubsection">
3201 <a name="i_add">'<tt>add</tt>' Instruction</a>
3202</div>
3203
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3204<div class="doc_text">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3205
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3206<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3207<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3208 &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 +0000[diff] [blame]3209 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3210 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3211 &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 +0000[diff] [blame]3212</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3213
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3214<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3215<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3216
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3217<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3218<p>The two arguments to the '<tt>add</tt>' instruction must
3219 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3220 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3221
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3222<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3223<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3224
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3225<p>If the sum has unsigned overflow, the result returned is the mathematical
3226 result modulo 2<sup>n</sup>, where n is the bit width of the result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3227
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3228<p>Because LLVM integers use a two's complement representation, this instruction
3229 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3230
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3231<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3232 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3233 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3234 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3235 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3236
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3237<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3238<pre>
3239 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3240</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3241
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3242</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3243
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3244<!-- _______________________________________________________________________ -->
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3245<div class="doc_subsubsection">
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3246 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
3247</div>
3248
3249<div class="doc_text">
3250
3251<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3252<pre>
3253 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3254</pre>
3255
3256<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3257<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3258
3259<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3260<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3261 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3262 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3263
3264<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3265<p>The value produced is the floating point sum of the two operands.</p>
3266
3267<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3268<pre>
3269 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3270</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3271
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3272</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3273
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3274<!-- _______________________________________________________________________ -->
3275<div class="doc_subsubsection">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3276 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
3277</div>
3278
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3279<div class="doc_text">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3280
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3281<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3282<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3283 &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 +0000[diff] [blame]3284 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3285 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3286 &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 +0000[diff] [blame]3287</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3288
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3289<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3290<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3291 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3292
3293<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3294 '<tt>neg</tt>' instruction present in most other intermediate
3295 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3296
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3297<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3298<p>The two arguments to the '<tt>sub</tt>' instruction must
3299 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3300 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3301
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3302<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3303<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3304
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3305<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3306 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3307 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3308
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3309<p>Because LLVM integers use a two's complement representation, this instruction
3310 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3311
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3312<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3313 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3314 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3315 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3316 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3317
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3318<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]3319<pre>
3320 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3321 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3322</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3323
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3324</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3325
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3326<!-- _______________________________________________________________________ -->
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3327<div class="doc_subsubsection">
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3328 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
3329</div>
3330
3331<div class="doc_text">
3332
3333<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3334<pre>
3335 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3336</pre>
3337
3338<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3339<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3340 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3341
3342<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3343 '<tt>fneg</tt>' instruction present in most other intermediate
3344 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3345
3346<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]3347<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3348 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3349 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3350
3351<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3352<p>The value produced is the floating point difference of the two operands.</p>
3353
3354<h5>Example:</h5>
3355<pre>
3356 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3357 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3358</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3359
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3360</div>
3361
3362<!-- _______________________________________________________________________ -->
3363<div class="doc_subsubsection">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3364 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
3365</div>
3366
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3367<div class="doc_text">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3368
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3369<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3370<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3371 &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 +0000[diff] [blame]3372 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3373 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3374 &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 +0000[diff] [blame]3375</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3376
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3377<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3378<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3379
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3380<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3381<p>The two arguments to the '<tt>mul</tt>' instruction must
3382 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3383 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3384
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3385<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3386<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3387
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3388<p>If the result of the multiplication has unsigned overflow, the result
3389 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3390 width of the result.</p>
3391
3392<p>Because LLVM integers use a two's complement representation, and the result
3393 is the same width as the operands, this instruction returns the correct
3394 result for both signed and unsigned integers. If a full product
3395 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3396 be sign-extended or zero-extended as appropriate to the width of the full
3397 product.</p>
3398
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3399<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3400 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3401 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3402 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3403 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3404
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3405<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3406<pre>
3407 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3408</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3409
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3410</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3411
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3412<!-- _______________________________________________________________________ -->
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3413<div class="doc_subsubsection">
3414 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
3415</div>
3416
3417<div class="doc_text">
3418
3419<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3420<pre>
3421 &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 +0000[diff] [blame]3422</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3423
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3424<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3425<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3426
3427<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3428<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3429 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3430 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3431
3432<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3433<p>The value produced is the floating point product of the two operands.</p>
3434
3435<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3436<pre>
3437 &lt;result&gt; = fmul float 4.0, %var <i>; yields {float}:result = 4.0 * %var</i>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3438</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3439
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3440</div>
3441
3442<!-- _______________________________________________________________________ -->
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3443<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
3444</a></div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3445
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3446<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3447
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3448<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3449<pre>
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3450 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3451 &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 +0000[diff] [blame]3452</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3453
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3454<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3455<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3456
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3457<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3458<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3459 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3460 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3461
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3462<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3463<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3464
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3465<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3466 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3467
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3468<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3469
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3470<p>If the <tt>exact</tt> keyword is present, the result value of the
3471 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3472 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3473
3474
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3475<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3476<pre>
3477 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3478</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3479
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3480</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3481
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3482<!-- _______________________________________________________________________ -->
3483<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
3484</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3485
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3486<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3487
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3488<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3489<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3490 &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 +0000[diff] [blame]3491 &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 +0000[diff] [blame]3492</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3493
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3494<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3495<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3496
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3497<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3498<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3499 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3500 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3501
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3502<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3503<p>The value produced is the signed integer quotient of the two operands rounded
3504 towards zero.</p>
3505
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3506<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3507 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3508
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3509<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3510 undefined behavior; this is a rare case, but can occur, for example, by doing
3511 a 32-bit division of -2147483648 by -1.</p>
3512
Dan Gohman9c5beed2009-07-22 00:04:19 +0000[diff] [blame]3513<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]3514 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohman38da9272010-07-11 00:08:34 +0000[diff] [blame]3515 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3516
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3517<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3518<pre>
3519 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3520</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3521
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3522</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3523
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3524<!-- _______________________________________________________________________ -->
3525<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3526Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3527
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3528<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3529
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3530<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3531<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3532 &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 +0000[diff] [blame]3533</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3534
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3535<h5>Overview:</h5>
3536<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3537
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3538<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3539<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3540 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3541 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3542
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3543<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3544<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3545
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3546<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3547<pre>
3548 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3549</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3550
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3551</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3552
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3553<!-- _______________________________________________________________________ -->
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3554<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3555</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3556
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3557<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3558
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3559<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3560<pre>
3561 &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 +0000[diff] [blame]3562</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3563
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3564<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3565<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3566 division of its two arguments.</p>
3567
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3568<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3569<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3570 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3571 values. Both arguments must have identical types.</p>
3572
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3573<h5>Semantics:</h5>
3574<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3575 This instruction always performs an unsigned division to get the
3576 remainder.</p>
3577
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3578<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3579 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3580
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3581<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3582
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3583<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3584<pre>
3585 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3586</pre>
3587
3588</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3589
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3590<!-- _______________________________________________________________________ -->
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3591<div class="doc_subsubsection">
3592 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
3593</div>
3594
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3595<div class="doc_text">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3596
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3597<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3598<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3599 &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 +0000[diff] [blame]3600</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3601
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3602<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3603<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3604 division of its two operands. This instruction can also take
3605 <a href="#t_vector">vector</a> versions of the values in which case the
3606 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]3607
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3608<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3609<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3610 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3611 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3612
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3613<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3614<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +0000[diff] [blame]3615 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3616 <i>modulo</i> operator (where the result is either zero or has the same sign
3617 as the divisor, <tt>op2</tt>) of a value.
3618 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3619 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3620 Math Forum</a>. For a table of how this is implemented in various languages,
3621 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3622 Wikipedia: modulo operation</a>.</p>
3623
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3624<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3625 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3626
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3627<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3628 Overflow also leads to undefined behavior; this is a rare case, but can
3629 occur, for example, by taking the remainder of a 32-bit division of
3630 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3631 lets srem be implemented using instructions that return both the result of
3632 the division and the remainder.)</p>
3633
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3634<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3635<pre>
3636 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3637</pre>
3638
3639</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3640
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3641<!-- _______________________________________________________________________ -->
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3642<div class="doc_subsubsection">
3643 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
3644
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3645<div class="doc_text">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3646
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3647<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3648<pre>
3649 &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 +0000[diff] [blame]3650</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3651
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3652<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3653<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3654 its two operands.</p>
3655
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3656<h5>Arguments:</h5>
3657<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3658 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3659 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3660
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3661<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3662<p>This instruction returns the <i>remainder</i> of a division. The remainder
3663 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3664
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3665<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3666<pre>
3667 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3668</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3669
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3670</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]3671
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]3672<!-- ======================================================================= -->
3673<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
3674Operations</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3675
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]3676<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3677
3678<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3679 program. They are generally very efficient instructions and can commonly be
3680 strength reduced from other instructions. They require two operands of the
3681 same type, execute an operation on them, and produce a single value. The
3682 resulting value is the same type as its operands.</p>
3683
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]3684</div>
3685
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3686<!-- _______________________________________________________________________ -->
3687<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
3688Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3689
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3690<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3691
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3692<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3693<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3694 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3695 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3696 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3697 &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 +0000[diff] [blame]3698</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3699
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3700<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3701<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3702 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3703
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3704<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3705<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3706 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3707 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3708
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3709<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3710<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3711 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3712 is (statically or dynamically) negative or equal to or larger than the number
3713 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3714 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3715 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3716
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3717<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3718 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
Chris Lattner66298c12011-02-09 16:44:44 +0000[diff] [blame]3719 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3720 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3721 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3722 they would if the shift were expressed as a mul instruction with the same
3723 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3724
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3725<h5>Example:</h5>
3726<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3727 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3728 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3729 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3730 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]3731 &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 +0000[diff] [blame]3732</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3733
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3734</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3735
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3736<!-- _______________________________________________________________________ -->
3737<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
3738Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3739
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3740<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3741
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3742<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3743<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3744 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3745 &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 +0000[diff] [blame]3746</pre>
3747
3748<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3749<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3750 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3751
3752<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3753<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3754 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3755 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3756
3757<h5>Semantics:</h5>
3758<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3759 significant bits of the result will be filled with zero bits after the shift.
3760 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3761 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3762 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3763 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3764
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3765<p>If the <tt>exact</tt> keyword is present, the result value of the
3766 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3767 shifted out are non-zero.</p>
3768
3769
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3770<h5>Example:</h5>
3771<pre>
3772 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3773 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3774 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3775 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3776 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]3777 &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 +0000[diff] [blame]3778</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3779
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3780</div>
3781
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]3782<!-- _______________________________________________________________________ -->
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3783<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
3784Instruction</a> </div>
3785<div class="doc_text">
3786
3787<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3788<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3789 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3790 &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 +0000[diff] [blame]3791</pre>
3792
3793<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3794<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3795 operand shifted to the right a specified number of bits with sign
3796 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3797
3798<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3799<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3800 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3801 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3802
3803<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3804<p>This instruction always performs an arithmetic shift right operation, The
3805 most significant bits of the result will be filled with the sign bit
3806 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3807 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3808 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3809 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3810
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3811<p>If the <tt>exact</tt> keyword is present, the result value of the
3812 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3813 shifted out are non-zero.</p>
3814
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3815<h5>Example:</h5>
3816<pre>
3817 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3818 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3819 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3820 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3821 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]3822 &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 +0000[diff] [blame]3823</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3824
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3825</div>
3826
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3827<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3828<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
3829Instruction</a> </div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3830
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3831<div class="doc_text">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3832
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3833<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3834<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3835 &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 +0000[diff] [blame]3836</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3837
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3838<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3839<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3840 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3841
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3842<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3843<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3844 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3845 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3846
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3847<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3848<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3849
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3850<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3851 <tbody>
3852 <tr>
3853 <td>In0</td>
3854 <td>In1</td>
3855 <td>Out</td>
3856 </tr>
3857 <tr>
3858 <td>0</td>
3859 <td>0</td>
3860 <td>0</td>
3861 </tr>
3862 <tr>
3863 <td>0</td>
3864 <td>1</td>
3865 <td>0</td>
3866 </tr>
3867 <tr>
3868 <td>1</td>
3869 <td>0</td>
3870 <td>0</td>
3871 </tr>
3872 <tr>
3873 <td>1</td>
3874 <td>1</td>
3875 <td>1</td>
3876 </tr>
3877 </tbody>
3878</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3879
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3880<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3881<pre>
3882 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3883 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
3884 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3885</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3886</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3887<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3888<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3889
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3890<div class="doc_text">
3891
3892<h5>Syntax:</h5>
3893<pre>
3894 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3895</pre>
3896
3897<h5>Overview:</h5>
3898<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
3899 two operands.</p>
3900
3901<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3902<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3903 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3904 values. Both arguments must have identical types.</p>
3905
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3906<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3907<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3908
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3909<table border="1" cellspacing="0" cellpadding="4">
3910 <tbody>
3911 <tr>
3912 <td>In0</td>
3913 <td>In1</td>
3914 <td>Out</td>
3915 </tr>
3916 <tr>
3917 <td>0</td>
3918 <td>0</td>
3919 <td>0</td>
3920 </tr>
3921 <tr>
3922 <td>0</td>
3923 <td>1</td>
3924 <td>1</td>
3925 </tr>
3926 <tr>
3927 <td>1</td>
3928 <td>0</td>
3929 <td>1</td>
3930 </tr>
3931 <tr>
3932 <td>1</td>
3933 <td>1</td>
3934 <td>1</td>
3935 </tr>
3936 </tbody>
3937</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3938
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3939<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3940<pre>
3941 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3942 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
3943 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3944</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3945
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3946</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3947
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3948<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3949<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
3950Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3951
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3952<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3953
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3954<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3955<pre>
3956 &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 +0000[diff] [blame]3957</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3958
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3959<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3960<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
3961 its two operands. The <tt>xor</tt> is used to implement the "one's
3962 complement" operation, which is the "~" operator in C.</p>
3963
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3964<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3965<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3966 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3967 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3968
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3969<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3970<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3971
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3972<table border="1" cellspacing="0" cellpadding="4">
3973 <tbody>
3974 <tr>
3975 <td>In0</td>
3976 <td>In1</td>
3977 <td>Out</td>
3978 </tr>
3979 <tr>
3980 <td>0</td>
3981 <td>0</td>
3982 <td>0</td>
3983 </tr>
3984 <tr>
3985 <td>0</td>
3986 <td>1</td>
3987 <td>1</td>
3988 </tr>
3989 <tr>
3990 <td>1</td>
3991 <td>0</td>
3992 <td>1</td>
3993 </tr>
3994 <tr>
3995 <td>1</td>
3996 <td>1</td>
3997 <td>0</td>
3998 </tr>
3999 </tbody>
4000</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4001
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4002<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4003<pre>
4004 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4005 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4006 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4007 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4008</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4009
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4010</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4011
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4012<!-- ======================================================================= -->
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4013<div class="doc_subsection">
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4014 <a name="vectorops">Vector Operations</a>
4015</div>
4016
4017<div class="doc_text">
4018
4019<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4020 target-independent manner. These instructions cover the element-access and
4021 vector-specific operations needed to process vectors effectively. While LLVM
4022 does directly support these vector operations, many sophisticated algorithms
4023 will want to use target-specific intrinsics to take full advantage of a
4024 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4025
4026</div>
4027
4028<!-- _______________________________________________________________________ -->
4029<div class="doc_subsubsection">
4030 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
4031</div>
4032
4033<div class="doc_text">
4034
4035<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4036<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4037 &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 +0000[diff] [blame]4038</pre>
4039
4040<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4041<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4042 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4043
4044
4045<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4046<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4047 of <a href="#t_vector">vector</a> type. The second operand is an index
4048 indicating the position from which to extract the element. The index may be
4049 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4050
4051<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4052<p>The result is a scalar of the same type as the element type of
4053 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4054 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4055 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4056
4057<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4058<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4059 &lt;result&gt; = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4060</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4061
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4062</div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4063
4064<!-- _______________________________________________________________________ -->
4065<div class="doc_subsubsection">
4066 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
4067</div>
4068
4069<div class="doc_text">
4070
4071<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4072<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +0000[diff] [blame]4073 &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 +0000[diff] [blame]4074</pre>
4075
4076<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4077<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4078 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4079
4080<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4081<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4082 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4083 whose type must equal the element type of the first operand. The third
4084 operand is an index indicating the position at which to insert the value.
4085 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4086
4087<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4088<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4089 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4090 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4091 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4092
4093<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4094<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4095 &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 +0000[diff] [blame]4096</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4097
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4098</div>
4099
4100<!-- _______________________________________________________________________ -->
4101<div class="doc_subsubsection">
4102 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
4103</div>
4104
4105<div class="doc_text">
4106
4107<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4108<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4109 &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 +0000[diff] [blame]4110</pre>
4111
4112<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4113<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4114 from two input vectors, returning a vector with the same element type as the
4115 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4116
4117<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4118<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4119 with types that match each other. The third argument is a shuffle mask whose
4120 element type is always 'i32'. The result of the instruction is a vector
4121 whose length is the same as the shuffle mask and whose element type is the
4122 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4123
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4124<p>The shuffle mask operand is required to be a constant vector with either
4125 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4126
4127<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4128<p>The elements of the two input vectors are numbered from left to right across
4129 both of the vectors. The shuffle mask operand specifies, for each element of
4130 the result vector, which element of the two input vectors the result element
4131 gets. The element selector may be undef (meaning "don't care") and the
4132 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4133
4134<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4135<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4136 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4137 &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 +0000[diff] [blame]4138 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4139 &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 +0000[diff] [blame]4140 &lt;result&gt; = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4141 &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 +0000[diff] [blame]4142 &lt;result&gt; = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4143 &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 +0000[diff] [blame]4144</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4145
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4146</div>
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]4147
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4148<!-- ======================================================================= -->
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4149<div class="doc_subsection">
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4150 <a name="aggregateops">Aggregate Operations</a>
4151</div>
4152
4153<div class="doc_text">
4154
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4155<p>LLVM supports several instructions for working with
4156 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4157
4158</div>
4159
4160<!-- _______________________________________________________________________ -->
4161<div class="doc_subsubsection">
4162 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
4163</div>
4164
4165<div class="doc_text">
4166
4167<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4168<pre>
4169 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4170</pre>
4171
4172<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4173<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4174 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4175
4176<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4177<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4178 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4179 <a href="#t_array">array</a> type. The operands are constant indices to
4180 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4181 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4182 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4183 <ul>
4184 <li>Since the value being indexed is not a pointer, the first index is
4185 omitted and assumed to be zero.</li>
4186 <li>At least one index must be specified.</li>
4187 <li>Not only struct indices but also array indices must be in
4188 bounds.</li>
4189 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4190
4191<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4192<p>The result is the value at the position in the aggregate specified by the
4193 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4194
4195<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4196<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4197 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4198</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4199
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4200</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4201
4202<!-- _______________________________________________________________________ -->
4203<div class="doc_subsubsection">
4204 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
4205</div>
4206
4207<div class="doc_text">
4208
4209<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4210<pre>
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]4211 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt; <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4212</pre>
4213
4214<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4215<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4216 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4217
4218<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4219<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4220 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4221 <a href="#t_array">array</a> type. The second operand is a first-class
4222 value to insert. The following operands are constant indices indicating
4223 the position at which to insert the value in a similar manner as indices in a
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4224 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4225 value to insert must have the same type as the value identified by the
4226 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4227
4228<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4229<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4230 that of <tt>val</tt> except that the value at the position specified by the
4231 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4232
4233<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4234<pre>
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]4235 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4236 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4237</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4238
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4239</div>
4240
4241
4242<!-- ======================================================================= -->
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4243<div class="doc_subsection">
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4244 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4245</div>
4246
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4247<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4248
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4249<p>A key design point of an SSA-based representation is how it represents
4250 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]4251 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4252 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4253
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4254</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4255
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4256<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4257<div class="doc_subsubsection">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4258 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
4259</div>
4260
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4261<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4262
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4263<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4264<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +0000[diff] [blame]4265 &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 +0000[diff] [blame]4266</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4267
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4268<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4269<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4270 currently executing function, to be automatically released when this function
4271 returns to its caller. The object is always allocated in the generic address
4272 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4273
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4274<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4275<p>The '<tt>alloca</tt>' instruction
4276 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4277 runtime stack, returning a pointer of the appropriate type to the program.
4278 If "NumElements" is specified, it is the number of elements allocated,
4279 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4280 specified, the value result of the allocation is guaranteed to be aligned to
4281 at least that boundary. If not specified, or if zero, the target can choose
4282 to align the allocation on any convenient boundary compatible with the
4283 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4284
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4285<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4286
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4287<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +0000[diff] [blame]4288<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4289 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4290 memory is automatically released when the function returns. The
4291 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4292 variables that must have an address available. When the function returns
4293 (either with the <tt><a href="#i_ret">ret</a></tt>
4294 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4295 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4296
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4297<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4298<pre>
Dan Gohman81e21672009-01-04 23:49:44 +0000[diff] [blame]4299 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4300 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4301 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4302 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4303</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4304
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4305</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4306
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4307<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4308<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
4309Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4310
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4311<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4312
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4313<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4314<pre>
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4315 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4316 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4317 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4318</pre>
4319
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4320<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4321<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4322
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4323<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4324<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4325 from which to load. The pointer must point to
4326 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4327 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4328 number or order of execution of this <tt>load</tt> with other <a
4329 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4330
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4331<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4332 operation (that is, the alignment of the memory address). A value of 0 or an
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4333 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4334 alignment for the target. It is the responsibility of the code emitter to
4335 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4336 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4337 produce less efficient code. An alignment of 1 is always safe.</p>
4338
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4339<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4340 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4341 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4342 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4343 and code generator that this load is not expected to be reused in the cache.
4344 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4345 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4346
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4347<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4348<p>The location of memory pointed to is loaded. If the value being loaded is of
4349 scalar type then the number of bytes read does not exceed the minimum number
4350 of bytes needed to hold all bits of the type. For example, loading an
4351 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4352 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4353 is undefined if the value was not originally written using a store of the
4354 same type.</p>
4355
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4356<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4357<pre>
4358 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4359 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4360 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4361</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4362
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4363</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4364
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4365<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4366<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
4367Instruction</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4368
Reid Spencer035ab572006-11-09 21:18:01 +0000[diff] [blame]4369<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4370
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4371<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4372<pre>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4373 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>
4374 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 +0000[diff] [blame]4375</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4376
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4377<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4378<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4379
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4380<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4381<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4382 and an address at which to store it. The type of the
4383 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4384 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4385 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4386 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4387 order of execution of this <tt>store</tt> with other <a
4388 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4389
4390<p>The optional constant "align" argument specifies the alignment of the
4391 operation (that is, the alignment of the memory address). A value of 0 or an
4392 omitted "align" argument means that the operation has the preferential
4393 alignment for the target. It is the responsibility of the code emitter to
4394 ensure that the alignment information is correct. Overestimating the
4395 alignment results in an undefined behavior. Underestimating the alignment may
4396 produce less efficient code. An alignment of 1 is always safe.</p>
4397
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4398<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4399 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4400 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4401 instruction tells the optimizer and code generator that this load is
4402 not expected to be reused in the cache. The code generator may
4403 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4404 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4405
4406
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4407<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4408<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4409 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4410 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4411 does not exceed the minimum number of bytes needed to hold all bits of the
4412 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4413 writing a value of a type like <tt>i20</tt> with a size that is not an
4414 integral number of bytes, it is unspecified what happens to the extra bits
4415 that do not belong to the type, but they will typically be overwritten.</p>
4416
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4417<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4418<pre>
4419 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]4420 store i32 3, i32* %ptr <i>; yields {void}</i>
4421 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4422</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4423
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]4424</div>
4425
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4426<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4427<div class="doc_subsubsection">
4428 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
4429</div>
4430
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4431<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4432
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4433<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4434<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4435 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]4436 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4437</pre>
4438
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4439<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4440<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4441 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4442 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4443
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4444<h5>Arguments:</h5>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4445<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]4446 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4447 elements of the aggregate object are indexed. The interpretation of each
4448 index is dependent on the type being indexed into. The first index always
4449 indexes the pointer value given as the first argument, the second index
4450 indexes a value of the type pointed to (not necessarily the value directly
4451 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4452 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4453 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4454 can never be pointers, since that would require loading the pointer before
4455 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4456
4457<p>The type of each index argument depends on the type it is indexing into.
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4458 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4459 integer <b>constants</b> are allowed. When indexing into an array, pointer
4460 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]4461 constant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4462
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4463<p>For example, let's consider a C code fragment and how it gets compiled to
4464 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4465
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4466<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4467struct RT {
4468 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4469 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4470 char C;
4471};
4472struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4473 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4474 double Y;
4475 struct RT Z;
4476};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4477
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4478int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4479 return &amp;s[1].Z.B[5][13];
4480}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4481</pre>
4482
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4483<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4484
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4485<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]4486%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4487%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4488
Dan Gohman4df605b2009-07-25 02:23:48 +0000[diff] [blame]4489define i32* @foo(%ST* %s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4490entry:
4491 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4492 ret i32* %reg
4493}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4494</pre>
4495
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4496<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4497<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4498 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4499 }</tt>' type, a structure. The second index indexes into the third element
4500 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4501 i8 }</tt>' type, another structure. The third index indexes into the second
4502 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4503 array. The two dimensions of the array are subscripted into, yielding an
4504 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4505 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4506
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4507<p>Note that it is perfectly legal to index partially through a structure,
4508 returning a pointer to an inner element. Because of this, the LLVM code for
4509 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4510
4511<pre>
Dan Gohman4df605b2009-07-25 02:23:48 +0000[diff] [blame]4512 define i32* @foo(%ST* %s) {
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4513 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4514 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4515 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4516 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4517 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4518 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4519 }
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]4520</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]4521
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]4522<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]4523 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4524 base pointer is not an <i>in bounds</i> address of an allocated object,
4525 or if any of the addresses that would be formed by successive addition of
4526 the offsets implied by the indices to the base address with infinitely
4527 precise arithmetic are not an <i>in bounds</i> address of that allocated
4528 object. The <i>in bounds</i> addresses for an allocated object are all
4529 the addresses that point into the object, plus the address one byte past
4530 the end.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]4531
4532<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4533 the base address with silently-wrapping two's complement arithmetic, and
4534 the result value of the <tt>getelementptr</tt> may be outside the object
4535 pointed to by the base pointer. The result value may not necessarily be
4536 used to access memory though, even if it happens to point into allocated
4537 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4538 section for more information.</p>
4539
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4540<p>The getelementptr instruction is often confusing. For some more insight into
4541 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4542
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4543<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4544<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4545 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4546 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4547 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]4548 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4549 <i>; yields i8*:eptr</i>
4550 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +0000[diff] [blame]4551 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +0000[diff] [blame]4552 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4553</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4554
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4555</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]4556
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4557<!-- ======================================================================= -->
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]4558<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4559</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4560
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4561<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4562
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]4563<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4564 which all take a single operand and a type. They perform various bit
4565 conversions on the operand.</p>
4566
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4567</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]4568
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]4569<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]4570<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4571 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
4572</div>
4573<div class="doc_text">
4574
4575<h5>Syntax:</h5>
4576<pre>
4577 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4578</pre>
4579
4580<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4581<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4582 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4583
4584<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4585<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
4586 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4587 of the same number of integers.
4588 The bit size of the <tt>value</tt> must be larger than
4589 the bit size of the destination type, <tt>ty2</tt>.
4590 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4591
4592<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4593<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4594 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4595 source size must be larger than the destination size, <tt>trunc</tt> cannot
4596 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4597
4598<h5>Example:</h5>
4599<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4600 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4601 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
4602 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
4603 %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 +0000[diff] [blame]4604</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4605
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4606</div>
4607
4608<!-- _______________________________________________________________________ -->
4609<div class="doc_subsubsection">
4610 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
4611</div>
4612<div class="doc_text">
4613
4614<h5>Syntax:</h5>
4615<pre>
4616 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4617</pre>
4618
4619<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4620<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4621 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4622
4623
4624<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]4625<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
4626 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4627 of the same number of integers.
4628 The bit size of the <tt>value</tt> must be smaller than
4629 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4630 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4631
4632<h5>Semantics:</h5>
4633<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4634 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4635
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]4636<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4637
4638<h5>Example:</h5>
4639<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4640 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]4641 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]4642 %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 +0000[diff] [blame]4643</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4644
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4645</div>
4646
4647<!-- _______________________________________________________________________ -->
4648<div class="doc_subsubsection">
4649 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
4650</div>
4651<div class="doc_text">
4652
4653<h5>Syntax:</h5>
4654<pre>
4655 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4656</pre>
4657
4658<h5>Overview:</h5>
4659<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4660
4661<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4662<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
4663 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4664 of the same number of integers.
4665 The bit size of the <tt>value</tt> must be smaller than
4666 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4667 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4668
4669<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4670<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4671 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4672 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4673
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]4674<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4675
4676<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4677<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4678 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]4679 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4680 %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 +0000[diff] [blame]4681</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4682
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4683</div>
4684
4685<!-- _______________________________________________________________________ -->
4686<div class="doc_subsubsection">
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4687 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
4688</div>
4689
4690<div class="doc_text">
4691
4692<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4693<pre>
4694 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4695</pre>
4696
4697<h5>Overview:</h5>
4698<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4699 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4700
4701<h5>Arguments:</h5>
4702<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4703 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4704 to cast it to. The size of <tt>value</tt> must be larger than the size of
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4705 <tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4706 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4707
4708<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4709<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4710 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4711 <a href="#t_floating">floating point</a> type. If the value cannot fit
4712 within the destination type, <tt>ty2</tt>, then the results are
4713 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4714
4715<h5>Example:</h5>
4716<pre>
4717 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4718 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4719</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4720
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4721</div>
4722
4723<!-- _______________________________________________________________________ -->
4724<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4725 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
4726</div>
4727<div class="doc_text">
4728
4729<h5>Syntax:</h5>
4730<pre>
4731 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4732</pre>
4733
4734<h5>Overview:</h5>
4735<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4736 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4737
4738<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4739<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4740 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4741 a <a href="#t_floating">floating point</a> type to cast it to. The source
4742 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4743
4744<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4745<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4746 <a href="#t_floating">floating point</a> type to a larger
4747 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4748 used to make a <i>no-op cast</i> because it always changes bits. Use
4749 <tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4750
4751<h5>Example:</h5>
4752<pre>
4753 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
4754 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
4755</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4756
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4757</div>
4758
4759<!-- _______________________________________________________________________ -->
4760<div class="doc_subsubsection">
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]4761 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4762</div>
4763<div class="doc_text">
4764
4765<h5>Syntax:</h5>
4766<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]4767 &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 +0000[diff] [blame]4768</pre>
4769
4770<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]4771<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4772 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4773
4774<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4775<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4776 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4777 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4778 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4779 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4780
4781<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4782<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4783 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4784 towards zero) unsigned integer value. If the value cannot fit
4785 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4786
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4787<h5>Example:</h5>
4788<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]4789 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]4790 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]4791 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4792</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4793
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4794</div>
4795
4796<!-- _______________________________________________________________________ -->
4797<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4798 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4799</div>
4800<div class="doc_text">
4801
4802<h5>Syntax:</h5>
4803<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4804 &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 +0000[diff] [blame]4805</pre>
4806
4807<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4808<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4809 <a href="#t_floating">floating point</a> <tt>value</tt> to
4810 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]4811
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]4812<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4813<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
4814 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4815 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4816 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4817 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]4818
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]4819<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4820<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4821 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4822 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
4823 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]4824
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]4825<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]4826<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]4827 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]4828 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]4829 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4830</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4831
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4832</div>
4833
4834<!-- _______________________________________________________________________ -->
4835<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4836 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4837</div>
4838<div class="doc_text">
4839
4840<h5>Syntax:</h5>
4841<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4842 &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 +0000[diff] [blame]4843</pre>
4844
4845<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4846<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4847 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4848
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4849<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]4850<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4851 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4852 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4853 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4854 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4855
4856<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4857<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4858 integer quantity and converts it to the corresponding floating point
4859 value. If the value cannot fit in the floating point value, the results are
4860 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4861
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4862<h5>Example:</h5>
4863<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4864 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]4865 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4866</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4867
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4868</div>
4869
4870<!-- _______________________________________________________________________ -->
4871<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4872 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4873</div>
4874<div class="doc_text">
4875
4876<h5>Syntax:</h5>
4877<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4878 &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 +0000[diff] [blame]4879</pre>
4880
4881<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4882<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
4883 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4884
4885<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]4886<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4887 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
4888 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
4889 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
4890 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4891
4892<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4893<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
4894 quantity and converts it to the corresponding floating point value. If the
4895 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4896
4897<h5>Example:</h5>
4898<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4899 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]4900 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4901</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4902
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4903</div>
4904
4905<!-- _______________________________________________________________________ -->
4906<div class="doc_subsubsection">
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4907 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
4908</div>
4909<div class="doc_text">
4910
4911<h5>Syntax:</h5>
4912<pre>
4913 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4914</pre>
4915
4916<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4917<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
4918 the integer type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4919
4920<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4921<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
4922 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
4923 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4924
4925<h5>Semantics:</h5>
4926<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4927 <tt>ty2</tt> by interpreting the pointer value as an integer and either
4928 truncating or zero extending that value to the size of the integer type. If
4929 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
4930 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
4931 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
4932 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4933
4934<h5>Example:</h5>
4935<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]4936 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
4937 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4938</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4939
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4940</div>
4941
4942<!-- _______________________________________________________________________ -->
4943<div class="doc_subsubsection">
4944 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
4945</div>
4946<div class="doc_text">
4947
4948<h5>Syntax:</h5>
4949<pre>
4950 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4951</pre>
4952
4953<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4954<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
4955 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4956
4957<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]4958<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4959 value to cast, and a type to cast it to, which must be a
4960 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4961
4962<h5>Semantics:</h5>
4963<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4964 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
4965 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
4966 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
4967 than the size of a pointer then a zero extension is done. If they are the
4968 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4969
4970<h5>Example:</h5>
4971<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]4972 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]4973 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
4974 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4975</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4976
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]4977</div>
4978
4979<!-- _______________________________________________________________________ -->
4980<div class="doc_subsubsection">
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]4981 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4982</div>
4983<div class="doc_text">
4984
4985<h5>Syntax:</h5>
4986<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]4987 &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 +0000[diff] [blame]4988</pre>
4989
4990<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]4991<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4992 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4993
4994<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4995<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
4996 non-aggregate first class value, and a type to cast it to, which must also be
4997 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
4998 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
4999 identical. If the source type is a pointer, the destination type must also be
5000 a pointer. This instruction supports bitwise conversion of vectors to
5001 integers and to vectors of other types (as long as they have the same
5002 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5003
5004<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5005<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5006 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5007 this conversion. The conversion is done as if the <tt>value</tt> had been
5008 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5009 be converted to other pointer types with this instruction. To convert
5010 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5011 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5012
5013<h5>Example:</h5>
5014<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5015 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5016 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5017 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5018</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5019
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5020</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5021
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5022<!-- ======================================================================= -->
5023<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5024
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5025<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5026
5027<p>The instructions in this category are the "miscellaneous" instructions, which
5028 defy better classification.</p>
5029
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5030</div>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5031
5032<!-- _______________________________________________________________________ -->
5033<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5034</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5035
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5036<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5037
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5038<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5039<pre>
5040 &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 +0000[diff] [blame]5041</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5042
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5043<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5044<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5045 boolean values based on comparison of its two integer, integer vector, or
5046 pointer operands.</p>
5047
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5048<h5>Arguments:</h5>
5049<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5050 the condition code indicating the kind of comparison to perform. It is not a
5051 value, just a keyword. The possible condition code are:</p>
5052
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5053<ol>
5054 <li><tt>eq</tt>: equal</li>
5055 <li><tt>ne</tt>: not equal </li>
5056 <li><tt>ugt</tt>: unsigned greater than</li>
5057 <li><tt>uge</tt>: unsigned greater or equal</li>
5058 <li><tt>ult</tt>: unsigned less than</li>
5059 <li><tt>ule</tt>: unsigned less or equal</li>
5060 <li><tt>sgt</tt>: signed greater than</li>
5061 <li><tt>sge</tt>: signed greater or equal</li>
5062 <li><tt>slt</tt>: signed less than</li>
5063 <li><tt>sle</tt>: signed less or equal</li>
5064</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5065
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]5066<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5067 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5068 typed. They must also be identical types.</p>
5069
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5070<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5071<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5072 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5073 either an <a href="#t_integer"><tt>i1</tt></a> or vector of <tt>i1</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5074 result, as follows:</p>
5075
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5076<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5077 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5078 <tt>false</tt> otherwise. No sign interpretation is necessary or
5079 performed.</li>
5080
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5081 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5082 <tt>false</tt> otherwise. No sign interpretation is necessary or
5083 performed.</li>
5084
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5085 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5086 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5087
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5088 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5089 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5090 to <tt>op2</tt>.</li>
5091
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5092 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5093 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5094
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5095 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5096 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5097
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5098 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5099 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5100
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5101 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5102 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5103 to <tt>op2</tt>.</li>
5104
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5105 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5106 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5107
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5108 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5109 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5110</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5111
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5112<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5113 values are compared as if they were integers.</p>
5114
5115<p>If the operands are integer vectors, then they are compared element by
5116 element. The result is an <tt>i1</tt> vector with the same number of elements
5117 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5118
5119<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5120<pre>
5121 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5122 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5123 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5124 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5125 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5126 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5127</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5128
5129<p>Note that the code generator does not yet support vector types with
5130 the <tt>icmp</tt> instruction.</p>
5131
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5132</div>
5133
5134<!-- _______________________________________________________________________ -->
5135<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5136</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5137
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5138<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5139
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5140<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5141<pre>
5142 &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 +0000[diff] [blame]5143</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5144
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5145<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5146<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5147 values based on comparison of its operands.</p>
5148
5149<p>If the operands are floating point scalars, then the result type is a boolean
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5150(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5151
5152<p>If the operands are floating point vectors, then the result type is a vector
5153 of boolean with the same number of elements as the operands being
5154 compared.</p>
5155
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5156<h5>Arguments:</h5>
5157<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5158 the condition code indicating the kind of comparison to perform. It is not a
5159 value, just a keyword. The possible condition code are:</p>
5160
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5161<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5162 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5163 <li><tt>oeq</tt>: ordered and equal</li>
5164 <li><tt>ogt</tt>: ordered and greater than </li>
5165 <li><tt>oge</tt>: ordered and greater than or equal</li>
5166 <li><tt>olt</tt>: ordered and less than </li>
5167 <li><tt>ole</tt>: ordered and less than or equal</li>
5168 <li><tt>one</tt>: ordered and not equal</li>
5169 <li><tt>ord</tt>: ordered (no nans)</li>
5170 <li><tt>ueq</tt>: unordered or equal</li>
5171 <li><tt>ugt</tt>: unordered or greater than </li>
5172 <li><tt>uge</tt>: unordered or greater than or equal</li>
5173 <li><tt>ult</tt>: unordered or less than </li>
5174 <li><tt>ule</tt>: unordered or less than or equal</li>
5175 <li><tt>une</tt>: unordered or not equal</li>
5176 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5177 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5178</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5179
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5180<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5181 <i>unordered</i> means that either operand may be a QNAN.</p>
5182
5183<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5184 a <a href="#t_floating">floating point</a> type or
5185 a <a href="#t_vector">vector</a> of floating point type. They must have
5186 identical types.</p>
5187
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5188<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]5189<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5190 according to the condition code given as <tt>cond</tt>. If the operands are
5191 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5192 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5193 follows:</p>
5194
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5195<ol>
5196 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5197
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5198 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5199 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5200
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5201 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5202 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5203
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5204 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5205 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5206
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5207 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5208 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5209
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5210 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5211 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5212
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5213 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5214 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5215
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5216 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5217
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5218 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5219 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5220
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5221 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5222 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5223
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5224 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5225 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5226
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5227 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5228 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5229
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5230 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5231 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5232
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5233 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5234 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5235
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5236 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5237
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5238 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5239</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5240
5241<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5242<pre>
5243 &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]5244 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5245 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5246 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5247</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5248
5249<p>Note that the code generator does not yet support vector types with
5250 the <tt>fcmp</tt> instruction.</p>
5251
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5252</div>
5253
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5254<!-- _______________________________________________________________________ -->
Nate Begemanac80ade2008-05-12 19:01:56 +0000[diff] [blame]5255<div class="doc_subsubsection">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5256 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
5257</div>
5258
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5259<div class="doc_text">
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5260
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5261<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5262<pre>
5263 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5264</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5265
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5266<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5267<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5268 SSA graph representing the function.</p>
5269
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5270<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5271<p>The type of the incoming values is specified with the first type field. After
5272 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5273 one pair for each predecessor basic block of the current block. Only values
5274 of <a href="#t_firstclass">first class</a> type may be used as the value
5275 arguments to the PHI node. Only labels may be used as the label
5276 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5277
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5278<p>There must be no non-phi instructions between the start of a basic block and
5279 the PHI instructions: i.e. PHI instructions must be first in a basic
5280 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5281
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5282<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5283 occur on the edge from the corresponding predecessor block to the current
5284 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5285 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +0000[diff] [blame]5286
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5287<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5288<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5289 specified by the pair corresponding to the predecessor basic block that
5290 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5291
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5292<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5293<pre>
5294Loop: ; Infinite loop that counts from 0 on up...
5295 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5296 %nextindvar = add i32 %indvar, 1
5297 br label %Loop
5298</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5299
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5300</div>
5301
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5302<!-- _______________________________________________________________________ -->
5303<div class="doc_subsubsection">
5304 <a name="i_select">'<tt>select</tt>' Instruction</a>
5305</div>
5306
5307<div class="doc_text">
5308
5309<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5310<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]5311 &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>
5312
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5313 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5314</pre>
5315
5316<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5317<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5318 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5319
5320
5321<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5322<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5323 values indicating the condition, and two values of the
5324 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5325 vectors and the condition is a scalar, then entire vectors are selected, not
5326 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5327
5328<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5329<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5330 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5331
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5332<p>If the condition is a vector of i1, then the value arguments must be vectors
5333 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5334
5335<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5336<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5337 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5338</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5339
5340<p>Note that the code generator does not yet support conditions
5341 with vector type.</p>
5342
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5343</div>
5344
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]5345<!-- _______________________________________________________________________ -->
5346<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5347 <a name="i_call">'<tt>call</tt>' Instruction</a>
5348</div>
5349
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5350<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5351
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5352<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5353<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]5354 &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 +0000[diff] [blame]5355</pre>
5356
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5357<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5358<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5359
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5360<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5361<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5362
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5363<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5364 <li>The optional "tail" marker indicates that the callee function does not
5365 access any allocas or varargs in the caller. Note that calls may be
5366 marked "tail" even if they do not occur before
5367 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5368 present, the function call is eligible for tail call optimization,
5369 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +0000[diff] [blame]5370 optimized into a jump</a>. The code generator may optimize calls marked
5371 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5372 sibling call optimization</a> when the caller and callee have
5373 matching signatures, or 2) forced tail call optimization when the
5374 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5375 <ul>
5376 <li>Caller and callee both have the calling
5377 convention <tt>fastcc</tt>.</li>
5378 <li>The call is in tail position (ret immediately follows call and ret
5379 uses value of call or is void).</li>
5380 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +0000[diff] [blame]5381 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5382 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5383 constraints are met.</a></li>
5384 </ul>
5385 </li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5386
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5387 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5388 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5389 defaults to using C calling conventions. The calling convention of the
5390 call must match the calling convention of the target function, or else the
5391 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5392
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5393 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5394 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5395 '<tt>inreg</tt>' attributes are valid here.</li>
5396
5397 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5398 type of the return value. Functions that return no value are marked
5399 <tt><a href="#t_void">void</a></tt>.</li>
5400
5401 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5402 being invoked. The argument types must match the types implied by this
5403 signature. This type can be omitted if the function is not varargs and if
5404 the function type does not return a pointer to a function.</li>
5405
5406 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5407 be invoked. In most cases, this is a direct function invocation, but
5408 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5409 to function value.</li>
5410
5411 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]5412 signature argument types and parameter attributes. All arguments must be
5413 of <a href="#t_firstclass">first class</a> type. If the function
5414 signature indicates the function accepts a variable number of arguments,
5415 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5416
5417 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5418 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5419 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5420</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5421
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5422<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5423<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5424 a specified function, with its incoming arguments bound to the specified
5425 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5426 function, control flow continues with the instruction after the function
5427 call, and the return value of the function is bound to the result
5428 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5429
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5430<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5431<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]5432 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]5433 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +0000[diff] [blame]5434 %X = tail call i32 @foo() <i>; yields i32</i>
5435 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5436 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]5437
5438 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5439 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]5440 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5441 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +0000[diff] [blame]5442 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +0000[diff] [blame]5443 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5444</pre>
5445
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]5446<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +0000[diff] [blame]5447standard C99 library as being the C99 library functions, and may perform
5448optimizations or generate code for them under that assumption. This is
5449something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5450freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]5451
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5452</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5453
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5454<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5455<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]5456 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5457</div>
5458
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5459<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5460
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5461<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5462<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5463 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5464</pre>
5465
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5466<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5467<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5468 the "variable argument" area of a function call. It is used to implement the
5469 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5470
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5471<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5472<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5473 argument. It returns a value of the specified argument type and increments
5474 the <tt>va_list</tt> to point to the next argument. The actual type
5475 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5476
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5477<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5478<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5479 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5480 to the next argument. For more information, see the variable argument
5481 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5482
5483<p>It is legal for this instruction to be called in a function which does not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5484 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5485 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5486
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5487<p><tt>va_arg</tt> is an LLVM instruction instead of
5488 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5489 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5490
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5491<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5492<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5493
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5494<p>Note that the code generator does not yet fully support va_arg on many
5495 targets. Also, it does not currently support va_arg with aggregate types on
5496 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]5497
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5498</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5499
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5500<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]5501<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
5502<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5503
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5504<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5505
5506<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5507 well known names and semantics and are required to follow certain
5508 restrictions. Overall, these intrinsics represent an extension mechanism for
5509 the LLVM language that does not require changing all of the transformations
5510 in LLVM when adding to the language (or the bitcode reader/writer, the
5511 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5512
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]5513<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5514 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5515 begin with this prefix. Intrinsic functions must always be external
5516 functions: you cannot define the body of intrinsic functions. Intrinsic
5517 functions may only be used in call or invoke instructions: it is illegal to
5518 take the address of an intrinsic function. Additionally, because intrinsic
5519 functions are part of the LLVM language, it is required if any are added that
5520 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5521
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5522<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5523 family of functions that perform the same operation but on different data
5524 types. Because LLVM can represent over 8 million different integer types,
5525 overloading is used commonly to allow an intrinsic function to operate on any
5526 integer type. One or more of the argument types or the result type can be
5527 overloaded to accept any integer type. Argument types may also be defined as
5528 exactly matching a previous argument's type or the result type. This allows
5529 an intrinsic function which accepts multiple arguments, but needs all of them
5530 to be of the same type, to only be overloaded with respect to a single
5531 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5532
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5533<p>Overloaded intrinsics will have the names of its overloaded argument types
5534 encoded into its function name, each preceded by a period. Only those types
5535 which are overloaded result in a name suffix. Arguments whose type is matched
5536 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5537 can take an integer of any width and returns an integer of exactly the same
5538 integer width. This leads to a family of functions such as
5539 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5540 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5541 suffix is required. Because the argument's type is matched against the return
5542 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]5543
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5544<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5545 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5546
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5547</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5548
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5549<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5550<div class="doc_subsection">
5551 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
5552</div>
5553
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5554<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5555
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5556<p>Variable argument support is defined in LLVM with
5557 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5558 intrinsic functions. These functions are related to the similarly named
5559 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5560
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5561<p>All of these functions operate on arguments that use a target-specific value
5562 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5563 not define what this type is, so all transformations should be prepared to
5564 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5565
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]5566<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5567 instruction and the variable argument handling intrinsic functions are
5568 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5569
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5570<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5571define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5572 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5573 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5574 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5575 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5576
5577 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5578 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5579
5580 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5581 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5582 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5583 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5584 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5585
5586 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5587 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5588 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5589}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5590
5591declare void @llvm.va_start(i8*)
5592declare void @llvm.va_copy(i8*, i8*)
5593declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5594</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5595
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5596</div>
5597
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5598<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5599<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5600 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5601</div>
5602
5603
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5604<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5605
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5606<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5607<pre>
5608 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5609</pre>
5610
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5611<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5612<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5613 for subsequent use by <tt><a href="#i_va_arg">va_arg</a></tt>.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5614
5615<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5616<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5617
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5618<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5619<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5620 macro available in C. In a target-dependent way, it initializes
5621 the <tt>va_list</tt> element to which the argument points, so that the next
5622 call to <tt>va_arg</tt> will produce the first variable argument passed to
5623 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5624 need to know the last argument of the function as the compiler can figure
5625 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5626
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5627</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5628
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5629<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5630<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5631 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5632</div>
5633
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5634<div class="doc_text">
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5635
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5636<h5>Syntax:</h5>
5637<pre>
5638 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5639</pre>
5640
5641<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5642<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5643 which has been initialized previously
5644 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5645 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5646
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5647<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5648<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5649
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5650<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5651<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5652 macro available in C. In a target-dependent way, it destroys
5653 the <tt>va_list</tt> element to which the argument points. Calls
5654 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5655 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5656 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5657
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5658</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5659
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5660<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5661<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5662 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5663</div>
5664
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5665<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5666
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5667<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5668<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]5669 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5670</pre>
5671
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5672<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5673<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5674 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5675
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5676<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5677<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5678 The second argument is a pointer to a <tt>va_list</tt> element to copy
5679 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5680
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5681<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5682<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5683 macro available in C. In a target-dependent way, it copies the
5684 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5685 element. This intrinsic is necessary because
5686 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5687 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5688
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5689</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5690
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5691<!-- ======================================================================= -->
5692<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5693 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
5694</div>
5695
5696<div class="doc_text">
5697
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5698<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]5699Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5700intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5701roots on the stack</a>, as well as garbage collector implementations that
5702require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5703barriers. Front-ends for type-safe garbage collected languages should generate
5704these intrinsics to make use of the LLVM garbage collectors. For more details,
5705see <a href="GarbageCollection.html">Accurate Garbage Collection with
5706LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]5707
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5708<p>The garbage collection intrinsics only operate on objects in the generic
5709 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]5710
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5711</div>
5712
5713<!-- _______________________________________________________________________ -->
5714<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5715 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5716</div>
5717
5718<div class="doc_text">
5719
5720<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5721<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5722 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5723</pre>
5724
5725<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]5726<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5727 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5728
5729<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5730<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5731 root pointer. The second pointer (which must be either a constant or a
5732 global value address) contains the meta-data to be associated with the
5733 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5734
5735<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]5736<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5737 location. At compile-time, the code generator generates information to allow
5738 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5739 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5740 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5741
5742</div>
5743
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5744<!-- _______________________________________________________________________ -->
5745<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5746 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5747</div>
5748
5749<div class="doc_text">
5750
5751<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5752<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5753 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5754</pre>
5755
5756<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5757<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5758 locations, allowing garbage collector implementations that require read
5759 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5760
5761<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]5762<p>The second argument is the address to read from, which should be an address
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5763 allocated from the garbage collector. The first object is a pointer to the
5764 start of the referenced object, if needed by the language runtime (otherwise
5765 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5766
5767<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5768<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5769 instruction, but may be replaced with substantially more complex code by the
5770 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5771 may only be used in a function which <a href="#gc">specifies a GC
5772 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5773
5774</div>
5775
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5776<!-- _______________________________________________________________________ -->
5777<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5778 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5779</div>
5780
5781<div class="doc_text">
5782
5783<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5784<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5785 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5786</pre>
5787
5788<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5789<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5790 locations, allowing garbage collector implementations that require write
5791 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5792
5793<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]5794<p>The first argument is the reference to store, the second is the start of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5795 object to store it to, and the third is the address of the field of Obj to
5796 store to. If the runtime does not require a pointer to the object, Obj may
5797 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5798
5799<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5800<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5801 instruction, but may be replaced with substantially more complex code by the
5802 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
5803 may only be used in a function which <a href="#gc">specifies a GC
5804 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5805
5806</div>
5807
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5808<!-- ======================================================================= -->
5809<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5810 <a name="int_codegen">Code Generator Intrinsics</a>
5811</div>
5812
5813<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5814
5815<p>These intrinsics are provided by LLVM to expose special features that may
5816 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5817
5818</div>
5819
5820<!-- _______________________________________________________________________ -->
5821<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5822 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5823</div>
5824
5825<div class="doc_text">
5826
5827<h5>Syntax:</h5>
5828<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]5829 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5830</pre>
5831
5832<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5833<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
5834 target-specific value indicating the return address of the current function
5835 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5836
5837<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5838<p>The argument to this intrinsic indicates which function to return the address
5839 for. Zero indicates the calling function, one indicates its caller, etc.
5840 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5841
5842<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5843<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
5844 indicating the return address of the specified call frame, or zero if it
5845 cannot be identified. The value returned by this intrinsic is likely to be
5846 incorrect or 0 for arguments other than zero, so it should only be used for
5847 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5848
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5849<p>Note that calling this intrinsic does not prevent function inlining or other
5850 aggressive transformations, so the value returned may not be that of the
5851 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5852
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5853</div>
5854
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5855<!-- _______________________________________________________________________ -->
5856<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5857 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5858</div>
5859
5860<div class="doc_text">
5861
5862<h5>Syntax:</h5>
5863<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]5864 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5865</pre>
5866
5867<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5868<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
5869 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5870
5871<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5872<p>The argument to this intrinsic indicates which function to return the frame
5873 pointer for. Zero indicates the calling function, one indicates its caller,
5874 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5875
5876<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5877<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
5878 indicating the frame address of the specified call frame, or zero if it
5879 cannot be identified. The value returned by this intrinsic is likely to be
5880 incorrect or 0 for arguments other than zero, so it should only be used for
5881 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5882
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5883<p>Note that calling this intrinsic does not prevent function inlining or other
5884 aggressive transformations, so the value returned may not be that of the
5885 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5886
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]5887</div>
5888
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]5889<!-- _______________________________________________________________________ -->
5890<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5891 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5892</div>
5893
5894<div class="doc_text">
5895
5896<h5>Syntax:</h5>
5897<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]5898 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5899</pre>
5900
5901<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5902<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
5903 of the function stack, for use
5904 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
5905 useful for implementing language features like scoped automatic variable
5906 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5907
5908<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5909<p>This intrinsic returns a opaque pointer value that can be passed
5910 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
5911 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
5912 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
5913 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
5914 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
5915 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5916
5917</div>
5918
5919<!-- _______________________________________________________________________ -->
5920<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5921 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5922</div>
5923
5924<div class="doc_text">
5925
5926<h5>Syntax:</h5>
5927<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]5928 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5929</pre>
5930
5931<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5932<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
5933 the function stack to the state it was in when the
5934 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
5935 executed. This is useful for implementing language features like scoped
5936 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5937
5938<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5939<p>See the description
5940 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5941
5942</div>
5943
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]5944<!-- _______________________________________________________________________ -->
5945<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5946 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]5947</div>
5948
5949<div class="doc_text">
5950
5951<h5>Syntax:</h5>
5952<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5953 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]5954</pre>
5955
5956<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5957<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
5958 insert a prefetch instruction if supported; otherwise, it is a noop.
5959 Prefetches have no effect on the behavior of the program but can change its
5960 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]5961
5962<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5963<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
5964 specifier determining if the fetch should be for a read (0) or write (1),
5965 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
5966 locality, to (3) - extremely local keep in cache. The <tt>rw</tt>
5967 and <tt>locality</tt> arguments must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]5968
5969<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5970<p>This intrinsic does not modify the behavior of the program. In particular,
5971 prefetches cannot trap and do not produce a value. On targets that support
5972 this intrinsic, the prefetch can provide hints to the processor cache for
5973 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]5974
5975</div>
5976
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]5977<!-- _______________________________________________________________________ -->
5978<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5979 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]5980</div>
5981
5982<div class="doc_text">
5983
5984<h5>Syntax:</h5>
5985<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5986 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]5987</pre>
5988
5989<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5990<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
5991 Counter (PC) in a region of code to simulators and other tools. The method
5992 is target specific, but it is expected that the marker will use exported
5993 symbols to transmit the PC of the marker. The marker makes no guarantees
5994 that it will remain with any specific instruction after optimizations. It is
5995 possible that the presence of a marker will inhibit optimizations. The
5996 intended use is to be inserted after optimizations to allow correlations of
5997 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]5998
5999<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6000<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6001
6002<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6003<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6004 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6005
6006</div>
6007
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6008<!-- _______________________________________________________________________ -->
6009<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6010 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6011</div>
6012
6013<div class="doc_text">
6014
6015<h5>Syntax:</h5>
6016<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]6017 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6018</pre>
6019
6020<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6021<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6022 counter register (or similar low latency, high accuracy clocks) on those
6023 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6024 should map to RPCC. As the backing counters overflow quickly (on the order
6025 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6026
6027<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6028<p>When directly supported, reading the cycle counter should not modify any
6029 memory. Implementations are allowed to either return a application specific
6030 value or a system wide value. On backends without support, this is lowered
6031 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6032
6033</div>
6034
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6035<!-- ======================================================================= -->
6036<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6037 <a name="int_libc">Standard C Library Intrinsics</a>
6038</div>
6039
6040<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6041
6042<p>LLVM provides intrinsics for a few important standard C library functions.
6043 These intrinsics allow source-language front-ends to pass information about
6044 the alignment of the pointer arguments to the code generator, providing
6045 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6046
6047</div>
6048
6049<!-- _______________________________________________________________________ -->
6050<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6051 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6052</div>
6053
6054<div class="doc_text">
6055
6056<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6057<p>This is an overloaded intrinsic. You can use <tt>llvm.memcpy</tt> on any
Mon P Wange88909b2010-04-07 06:35:53 +0000[diff] [blame]6058 integer bit width and for different address spaces. Not all targets support
6059 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6060
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6061<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6062 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6063 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6064 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6065 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6066</pre>
6067
6068<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6069<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6070 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6071
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6072<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6073 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6074 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6075
6076<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6077
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6078<p>The first argument is a pointer to the destination, the second is a pointer
6079 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6080 number of bytes to copy, the fourth argument is the alignment of the
6081 source and destination locations, and the fifth is a boolean indicating a
6082 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6083
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6084<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6085 then the caller guarantees that both the source and destination pointers are
6086 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6087
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6088<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6089 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6090 The detailed access behavior is not very cleanly specified and it is unwise
6091 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6092
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6093<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6094
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6095<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6096 source location to the destination location, which are not allowed to
6097 overlap. It copies "len" bytes of memory over. If the argument is known to
6098 be aligned to some boundary, this can be specified as the fourth argument,
6099 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6100
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6101</div>
6102
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6103<!-- _______________________________________________________________________ -->
6104<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6105 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6106</div>
6107
6108<div class="doc_text">
6109
6110<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6111<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
Mon P Wange88909b2010-04-07 06:35:53 +0000[diff] [blame]6112 width and for different address space. Not all targets support all bit
6113 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6114
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6115<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6116 declare void @llvm.memmove.p0i8.p0i8.i32(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6117 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6118 declare void @llvm.memmove.p0i8.p0i8.i64(i8* &lt;dest&gt;, i8* &lt;src&gt;,
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6119 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6120</pre>
6121
6122<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6123<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6124 source location to the destination location. It is similar to the
6125 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6126 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6127
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6128<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6129 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6130 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6131
6132<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6133
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6134<p>The first argument is a pointer to the destination, the second is a pointer
6135 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6136 number of bytes to copy, the fourth argument is the alignment of the
6137 source and destination locations, and the fifth is a boolean indicating a
6138 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6139
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6140<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6141 then the caller guarantees that the source and destination pointers are
6142 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6143
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6144<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6145 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6146 The detailed access behavior is not very cleanly specified and it is unwise
6147 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6148
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6149<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6150
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6151<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6152 source location to the destination location, which may overlap. It copies
6153 "len" bytes of memory over. If the argument is known to be aligned to some
6154 boundary, this can be specified as the fourth argument, otherwise it should
6155 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6156
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6157</div>
6158
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6159<!-- _______________________________________________________________________ -->
6160<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6161 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6162</div>
6163
6164<div class="doc_text">
6165
6166<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6167<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6168 width and for different address spaces. However, not all targets support all
6169 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6170
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6171<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6172 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6173 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6174 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6175 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6176</pre>
6177
6178<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6179<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6180 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6181
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6182<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6183 intrinsic does not return a value and takes extra alignment/volatile
6184 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6185
6186<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6187<p>The first argument is a pointer to the destination to fill, the second is the
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6188 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6189 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6190 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6191
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6192<p>If the call to this intrinsic has an alignment value that is not 0 or 1,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6193 then the caller guarantees that the destination pointer is aligned to that
6194 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6195
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6196<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6197 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6198 The detailed access behavior is not very cleanly specified and it is unwise
6199 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6200
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6201<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6202<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6203 at the destination location. If the argument is known to be aligned to some
6204 boundary, this can be specified as the fourth argument, otherwise it should
6205 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6206
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6207</div>
6208
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]6209<!-- _______________________________________________________________________ -->
6210<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6211 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6212</div>
6213
6214<div class="doc_text">
6215
6216<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6217<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6218 floating point or vector of floating point type. Not all targets support all
6219 types however.</p>
6220
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6221<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]6222 declare float @llvm.sqrt.f32(float %Val)
6223 declare double @llvm.sqrt.f64(double %Val)
6224 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6225 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6226 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6227</pre>
6228
6229<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6230<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6231 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6232 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6233 behavior for negative numbers other than -0.0 (which allows for better
6234 optimization, because there is no need to worry about errno being
6235 set). <tt>llvm.sqrt(-0.0)</tt> is defined to return -0.0 like IEEE sqrt.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6236
6237<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6238<p>The argument and return value are floating point numbers of the same
6239 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6240
6241<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6242<p>This function returns the sqrt of the specified operand if it is a
6243 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6244
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6245</div>
6246
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6247<!-- _______________________________________________________________________ -->
6248<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6249 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6250</div>
6251
6252<div class="doc_text">
6253
6254<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6255<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6256 floating point or vector of floating point type. Not all targets support all
6257 types however.</p>
6258
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6259<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]6260 declare float @llvm.powi.f32(float %Val, i32 %power)
6261 declare double @llvm.powi.f64(double %Val, i32 %power)
6262 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6263 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6264 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6265</pre>
6266
6267<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6268<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6269 specified (positive or negative) power. The order of evaluation of
6270 multiplications is not defined. When a vector of floating point type is
6271 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6272
6273<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6274<p>The second argument is an integer power, and the first is a value to raise to
6275 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6276
6277<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6278<p>This function returns the first value raised to the second power with an
6279 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6280
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6281</div>
6282
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6283<!-- _______________________________________________________________________ -->
6284<div class="doc_subsubsection">
6285 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
6286</div>
6287
6288<div class="doc_text">
6289
6290<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6291<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6292 floating point or vector of floating point type. Not all targets support all
6293 types however.</p>
6294
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6295<pre>
6296 declare float @llvm.sin.f32(float %Val)
6297 declare double @llvm.sin.f64(double %Val)
6298 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6299 declare fp128 @llvm.sin.f128(fp128 %Val)
6300 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6301</pre>
6302
6303<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6304<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6305
6306<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6307<p>The argument and return value are floating point numbers of the same
6308 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6309
6310<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6311<p>This function returns the sine of the specified operand, returning the same
6312 values as the libm <tt>sin</tt> functions would, and handles error conditions
6313 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6314
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6315</div>
6316
6317<!-- _______________________________________________________________________ -->
6318<div class="doc_subsubsection">
6319 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
6320</div>
6321
6322<div class="doc_text">
6323
6324<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6325<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6326 floating point or vector of floating point type. Not all targets support all
6327 types however.</p>
6328
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6329<pre>
6330 declare float @llvm.cos.f32(float %Val)
6331 declare double @llvm.cos.f64(double %Val)
6332 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6333 declare fp128 @llvm.cos.f128(fp128 %Val)
6334 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6335</pre>
6336
6337<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6338<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6339
6340<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6341<p>The argument and return value are floating point numbers of the same
6342 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6343
6344<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6345<p>This function returns the cosine of the specified operand, returning the same
6346 values as the libm <tt>cos</tt> functions would, and handles error conditions
6347 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6348
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6349</div>
6350
6351<!-- _______________________________________________________________________ -->
6352<div class="doc_subsubsection">
6353 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
6354</div>
6355
6356<div class="doc_text">
6357
6358<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6359<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6360 floating point or vector of floating point type. Not all targets support all
6361 types however.</p>
6362
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6363<pre>
6364 declare float @llvm.pow.f32(float %Val, float %Power)
6365 declare double @llvm.pow.f64(double %Val, double %Power)
6366 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6367 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6368 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6369</pre>
6370
6371<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6372<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6373 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6374
6375<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6376<p>The second argument is a floating point power, and the first is a value to
6377 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6378
6379<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6380<p>This function returns the first value raised to the second power, returning
6381 the same values as the libm <tt>pow</tt> functions would, and handles error
6382 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6383
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6384</div>
6385
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6386<!-- ======================================================================= -->
6387<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6388 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6389</div>
6390
6391<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6392
6393<p>LLVM provides intrinsics for a few important bit manipulation operations.
6394 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6395
6396</div>
6397
6398<!-- _______________________________________________________________________ -->
6399<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6400 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6401</div>
6402
6403<div class="doc_text">
6404
6405<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6406<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6407 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6408
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6409<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6410 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6411 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6412 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6413</pre>
6414
6415<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6416<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6417 values with an even number of bytes (positive multiple of 16 bits). These
6418 are useful for performing operations on data that is not in the target's
6419 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6420
6421<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6422<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6423 and low byte of the input i16 swapped. Similarly,
6424 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6425 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6426 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6427 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6428 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6429 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6430
6431</div>
6432
6433<!-- _______________________________________________________________________ -->
6434<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]6435 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6436</div>
6437
6438<div class="doc_text">
6439
6440<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6441<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6442 width. Not all targets support all bit widths however.</p>
6443
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6444<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6445 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6446 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6447 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6448 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6449 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6450</pre>
6451
6452<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6453<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6454 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6455
6456<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6457<p>The only argument is the value to be counted. The argument may be of any
6458 integer type. The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6459
6460<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6461<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6462
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6463</div>
6464
6465<!-- _______________________________________________________________________ -->
6466<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]6467 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6468</div>
6469
6470<div class="doc_text">
6471
6472<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6473<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
6474 integer bit width. Not all targets support all bit widths however.</p>
6475
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6476<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6477 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6478 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6479 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6480 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6481 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6482</pre>
6483
6484<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6485<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6486 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6487
6488<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6489<p>The only argument is the value to be counted. The argument may be of any
6490 integer type. The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6491
6492<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6493<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
6494 zeros in a variable. If the src == 0 then the result is the size in bits of
6495 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6496
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6497</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]6498
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6499<!-- _______________________________________________________________________ -->
6500<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]6501 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6502</div>
6503
6504<div class="doc_text">
6505
6506<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6507<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
6508 integer bit width. Not all targets support all bit widths however.</p>
6509
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6510<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6511 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6512 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6513 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6514 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6515 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6516</pre>
6517
6518<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6519<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6520 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6521
6522<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6523<p>The only argument is the value to be counted. The argument may be of any
6524 integer type. The return type must match the argument type.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6525
6526<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6527<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
6528 zeros in a variable. If the src == 0 then the result is the size in bits of
6529 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6530
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6531</div>
6532
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6533<!-- ======================================================================= -->
6534<div class="doc_subsection">
6535 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
6536</div>
6537
6538<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6539
6540<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6541
6542</div>
6543
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6544<!-- _______________________________________________________________________ -->
6545<div class="doc_subsubsection">
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6546 <a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6547</div>
6548
6549<div class="doc_text">
6550
6551<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6552<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6553 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6554
6555<pre>
6556 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6557 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6558 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6559</pre>
6560
6561<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6562<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6563 a signed addition of the two arguments, and indicate whether an overflow
6564 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6565
6566<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6567<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6568 be of integer types of any bit width, but they must have the same bit
6569 width. The second element of the result structure must be of
6570 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6571 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6572
6573<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6574<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6575 a signed addition of the two variables. They return a structure &mdash; the
6576 first element of which is the signed summation, and the second element of
6577 which is a bit specifying if the signed summation resulted in an
6578 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6579
6580<h5>Examples:</h5>
6581<pre>
6582 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6583 %sum = extractvalue {i32, i1} %res, 0
6584 %obit = extractvalue {i32, i1} %res, 1
6585 br i1 %obit, label %overflow, label %normal
6586</pre>
6587
6588</div>
6589
6590<!-- _______________________________________________________________________ -->
6591<div class="doc_subsubsection">
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6592 <a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6593</div>
6594
6595<div class="doc_text">
6596
6597<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6598<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6599 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6600
6601<pre>
6602 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6603 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6604 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6605</pre>
6606
6607<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6608<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6609 an unsigned addition of the two arguments, and indicate whether a carry
6610 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6611
6612<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6613<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6614 be of integer types of any bit width, but they must have the same bit
6615 width. The second element of the result structure must be of
6616 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6617 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6618
6619<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6620<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6621 an unsigned addition of the two arguments. They return a structure &mdash;
6622 the first element of which is the sum, and the second element of which is a
6623 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6624
6625<h5>Examples:</h5>
6626<pre>
6627 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6628 %sum = extractvalue {i32, i1} %res, 0
6629 %obit = extractvalue {i32, i1} %res, 1
6630 br i1 %obit, label %carry, label %normal
6631</pre>
6632
6633</div>
6634
6635<!-- _______________________________________________________________________ -->
6636<div class="doc_subsubsection">
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6637 <a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6638</div>
6639
6640<div class="doc_text">
6641
6642<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6643<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6644 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6645
6646<pre>
6647 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6648 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6649 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6650</pre>
6651
6652<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6653<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6654 a signed subtraction of the two arguments, and indicate whether an overflow
6655 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6656
6657<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6658<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6659 be of integer types of any bit width, but they must have the same bit
6660 width. The second element of the result structure must be of
6661 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6662 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6663
6664<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6665<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6666 a signed subtraction of the two arguments. They return a structure &mdash;
6667 the first element of which is the subtraction, and the second element of
6668 which is a bit specifying if the signed subtraction resulted in an
6669 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6670
6671<h5>Examples:</h5>
6672<pre>
6673 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6674 %sum = extractvalue {i32, i1} %res, 0
6675 %obit = extractvalue {i32, i1} %res, 1
6676 br i1 %obit, label %overflow, label %normal
6677</pre>
6678
6679</div>
6680
6681<!-- _______________________________________________________________________ -->
6682<div class="doc_subsubsection">
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6683 <a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6684</div>
6685
6686<div class="doc_text">
6687
6688<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6689<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6690 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6691
6692<pre>
6693 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
6694 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6695 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
6696</pre>
6697
6698<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6699<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6700 an unsigned subtraction of the two arguments, and indicate whether an
6701 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6702
6703<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6704<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6705 be of integer types of any bit width, but they must have the same bit
6706 width. The second element of the result structure must be of
6707 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6708 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6709
6710<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6711<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6712 an unsigned subtraction of the two arguments. They return a structure &mdash;
6713 the first element of which is the subtraction, and the second element of
6714 which is a bit specifying if the unsigned subtraction resulted in an
6715 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6716
6717<h5>Examples:</h5>
6718<pre>
6719 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
6720 %sum = extractvalue {i32, i1} %res, 0
6721 %obit = extractvalue {i32, i1} %res, 1
6722 br i1 %obit, label %overflow, label %normal
6723</pre>
6724
6725</div>
6726
6727<!-- _______________________________________________________________________ -->
6728<div class="doc_subsubsection">
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6729 <a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6730</div>
6731
6732<div class="doc_text">
6733
6734<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6735<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6736 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6737
6738<pre>
6739 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
6740 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6741 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
6742</pre>
6743
6744<h5>Overview:</h5>
6745
6746<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6747 a signed multiplication of the two arguments, and indicate whether an
6748 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6749
6750<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6751<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6752 be of integer types of any bit width, but they must have the same bit
6753 width. The second element of the result structure must be of
6754 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6755 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6756
6757<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6758<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6759 a signed multiplication of the two arguments. They return a structure &mdash;
6760 the first element of which is the multiplication, and the second element of
6761 which is a bit specifying if the signed multiplication resulted in an
6762 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6763
6764<h5>Examples:</h5>
6765<pre>
6766 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
6767 %sum = extractvalue {i32, i1} %res, 0
6768 %obit = extractvalue {i32, i1} %res, 1
6769 br i1 %obit, label %overflow, label %normal
6770</pre>
6771
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]6772</div>
6773
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6774<!-- _______________________________________________________________________ -->
6775<div class="doc_subsubsection">
6776 <a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
6777</div>
6778
6779<div class="doc_text">
6780
6781<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6782<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6783 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6784
6785<pre>
6786 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
6787 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6788 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
6789</pre>
6790
6791<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6792<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6793 a unsigned multiplication of the two arguments, and indicate whether an
6794 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6795
6796<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6797<p>The arguments (%a and %b) and the first element of the result structure may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6798 be of integer types of any bit width, but they must have the same bit
6799 width. The second element of the result structure must be of
6800 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6801 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6802
6803<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6804<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6805 an unsigned multiplication of the two arguments. They return a structure
6806 &mdash; the first element of which is the multiplication, and the second
6807 element of which is a bit specifying if the unsigned multiplication resulted
6808 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]6809
6810<h5>Examples:</h5>
6811<pre>
6812 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
6813 %sum = extractvalue {i32, i1} %res, 0
6814 %obit = extractvalue {i32, i1} %res, 1
6815 br i1 %obit, label %overflow, label %normal
6816</pre>
6817
6818</div>
6819
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6820<!-- ======================================================================= -->
6821<div class="doc_subsection">
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]6822 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
6823</div>
6824
6825<div class="doc_text">
6826
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]6827<p>Half precision floating point is a storage-only format. This means that it is
6828 a dense encoding (in memory) but does not support computation in the
6829 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]6830
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]6831<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]6832 value as an i16, then convert it to float with <a
6833 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
6834 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]6835 double etc). To store the value back to memory, it is first converted to
6836 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]6837 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
6838 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]6839</div>
6840
6841<!-- _______________________________________________________________________ -->
6842<div class="doc_subsubsection">
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]6843 <a name="int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]6844</div>
6845
6846<div class="doc_text">
6847
6848<h5>Syntax:</h5>
6849<pre>
6850 declare i16 @llvm.convert.to.fp16(f32 %a)
6851</pre>
6852
6853<h5>Overview:</h5>
6854<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6855 a conversion from single precision floating point format to half precision
6856 floating point format.</p>
6857
6858<h5>Arguments:</h5>
6859<p>The intrinsic function contains single argument - the value to be
6860 converted.</p>
6861
6862<h5>Semantics:</h5>
6863<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
6864 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]6865 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]6866 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]6867
6868<h5>Examples:</h5>
6869<pre>
6870 %res = call i16 @llvm.convert.to.fp16(f32 %a)
6871 store i16 %res, i16* @x, align 2
6872</pre>
6873
6874</div>
6875
6876<!-- _______________________________________________________________________ -->
6877<div class="doc_subsubsection">
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]6878 <a name="int_convert_from_fp16">'<tt>llvm.convert.from.fp16</tt>' Intrinsic</a>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]6879</div>
6880
6881<div class="doc_text">
6882
6883<h5>Syntax:</h5>
6884<pre>
6885 declare f32 @llvm.convert.from.fp16(i16 %a)
6886</pre>
6887
6888<h5>Overview:</h5>
6889<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
6890 a conversion from half precision floating point format to single precision
6891 floating point format.</p>
6892
6893<h5>Arguments:</h5>
6894<p>The intrinsic function contains single argument - the value to be
6895 converted.</p>
6896
6897<h5>Semantics:</h5>
6898<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]6899 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]6900 precision floating point format. The input half-float value is represented by
6901 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]6902
6903<h5>Examples:</h5>
6904<pre>
6905 %a = load i16* @x, align 2
6906 %res = call f32 @llvm.convert.from.fp16(i16 %a)
6907</pre>
6908
6909</div>
6910
6911<!-- ======================================================================= -->
6912<div class="doc_subsection">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6913 <a name="int_debugger">Debugger Intrinsics</a>
6914</div>
6915
6916<div class="doc_text">
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6917
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6918<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
6919 prefix), are described in
6920 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
6921 Level Debugging</a> document.</p>
6922
6923</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6924
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]6925<!-- ======================================================================= -->
6926<div class="doc_subsection">
6927 <a name="int_eh">Exception Handling Intrinsics</a>
6928</div>
6929
6930<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6931
6932<p>The LLVM exception handling intrinsics (which all start with
6933 <tt>llvm.eh.</tt> prefix), are described in
6934 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
6935 Handling</a> document.</p>
6936
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]6937</div>
6938
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]6939<!-- ======================================================================= -->
6940<div class="doc_subsection">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]6941 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6942</div>
6943
6944<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6945
6946<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]6947 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
6948 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6949 function pointer lacking the nest parameter - the caller does not need to
6950 provide a value for it. Instead, the value to use is stored in advance in a
6951 "trampoline", a block of memory usually allocated on the stack, which also
6952 contains code to splice the nest value into the argument list. This is used
6953 to implement the GCC nested function address extension.</p>
6954
6955<p>For example, if the function is
6956 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
6957 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
6958 follows:</p>
6959
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]6960<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]6961 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
6962 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]6963 %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 +0000[diff] [blame]6964 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6965</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6966
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]6967<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
6968 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6969
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6970</div>
6971
6972<!-- _______________________________________________________________________ -->
6973<div class="doc_subsubsection">
6974 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
6975</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6976
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6977<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6978
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6979<h5>Syntax:</h5>
6980<pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6981 declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6982</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6983
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6984<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6985<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
6986 function pointer suitable for executing it.</p>
6987
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6988<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6989<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
6990 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
6991 sufficiently aligned block of memory; this memory is written to by the
6992 intrinsic. Note that the size and the alignment are target-specific - LLVM
6993 currently provides no portable way of determining them, so a front-end that
6994 generates this intrinsic needs to have some target-specific knowledge.
6995 The <tt>func</tt> argument must hold a function bitcast to
6996 an <tt>i8*</tt>.</p>
6997
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]6998<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6999<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7000 dependent code, turning it into a function. A pointer to this function is
7001 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7002 function pointer type</a> before being called. The new function's signature
7003 is the same as that of <tt>func</tt> with any arguments marked with
7004 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7005 is allowed, and it must be of pointer type. Calling the new function is
7006 equivalent to calling <tt>func</tt> with the same argument list, but
7007 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7008 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7009 by <tt>tramp</tt> is modified, then the effect of any later call to the
7010 returned function pointer is undefined.</p>
7011
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7012</div>
7013
7014<!-- ======================================================================= -->
7015<div class="doc_subsection">
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7016 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
7017</div>
7018
7019<div class="doc_text">
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7020
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7021<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7022 hardware constructs for atomic operations and memory synchronization. This
7023 provides an interface to the hardware, not an interface to the programmer. It
7024 is aimed at a low enough level to allow any programming models or APIs
7025 (Application Programming Interfaces) which need atomic behaviors to map
7026 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7027 hardware provides a "universal IR" for source languages, it also provides a
7028 starting point for developing a "universal" atomic operation and
7029 synchronization IR.</p>
7030
7031<p>These do <em>not</em> form an API such as high-level threading libraries,
7032 software transaction memory systems, atomic primitives, and intrinsic
7033 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7034 application libraries. The hardware interface provided by LLVM should allow
7035 a clean implementation of all of these APIs and parallel programming models.
7036 No one model or paradigm should be selected above others unless the hardware
7037 itself ubiquitously does so.</p>
7038
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7039</div>
7040
7041<!-- _______________________________________________________________________ -->
7042<div class="doc_subsubsection">
7043 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
7044</div>
7045<div class="doc_text">
7046<h5>Syntax:</h5>
7047<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7048 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 +0000[diff] [blame]7049</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7050
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7051<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7052<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7053 specific pairs of memory access types.</p>
7054
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7055<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7056<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7057 The first four arguments enables a specific barrier as listed below. The
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7058 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7059 memory.</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7060
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7061<ul>
7062 <li><tt>ll</tt>: load-load barrier</li>
7063 <li><tt>ls</tt>: load-store barrier</li>
7064 <li><tt>sl</tt>: store-load barrier</li>
7065 <li><tt>ss</tt>: store-store barrier</li>
7066 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7067</ul>
7068
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7069<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7070<p>This intrinsic causes the system to enforce some ordering constraints upon
7071 the loads and stores of the program. This barrier does not
7072 indicate <em>when</em> any events will occur, it only enforces
7073 an <em>order</em> in which they occur. For any of the specified pairs of load
7074 and store operations (f.ex. load-load, or store-load), all of the first
7075 operations preceding the barrier will complete before any of the second
7076 operations succeeding the barrier begin. Specifically the semantics for each
7077 pairing is as follows:</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7078
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7079<ul>
7080 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7081 after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7082 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7083 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7084 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7085 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7086 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7087 load after the barrier begins.</li>
7088</ul>
7089
7090<p>These semantics are applied with a logical "and" behavior when more than one
7091 is enabled in a single memory barrier intrinsic.</p>
7092
7093<p>Backends may implement stronger barriers than those requested when they do
7094 not support as fine grained a barrier as requested. Some architectures do
7095 not need all types of barriers and on such architectures, these become
7096 noops.</p>
7097
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7098<h5>Example:</h5>
7099<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7100%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7101%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7102 store i32 4, %ptr
7103
7104%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7105 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false)
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7106 <i>; guarantee the above finishes</i>
7107 store i32 8, %ptr <i>; before this begins</i>
7108</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7109
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7110</div>
7111
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7112<!-- _______________________________________________________________________ -->
7113<div class="doc_subsubsection">
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7114 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7115</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7116
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7117<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7118
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7119<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7120<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7121 any integer bit width and for different address spaces. Not all targets
7122 support all bit widths however.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7123
7124<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7125 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7126 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7127 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7128 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 +0000[diff] [blame]7129</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7130
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7131<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7132<p>This loads a value in memory and compares it to a given value. If they are
7133 equal, it stores a new value into the memory.</p>
7134
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7135<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7136<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7137 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7138 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7139 this integer type. While any bit width integer may be used, targets may only
7140 lower representations they support in hardware.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7141
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7142<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7143<p>This entire intrinsic must be executed atomically. It first loads the value
7144 in memory pointed to by <tt>ptr</tt> and compares it with the
7145 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7146 memory. The loaded value is yielded in all cases. This provides the
7147 equivalent of an atomic compare-and-swap operation within the SSA
7148 framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7149
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7150<h5>Examples:</h5>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7151<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7152%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7153%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7154 store i32 4, %ptr
7155
7156%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7157%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7158 <i>; yields {i32}:result1 = 4</i>
7159%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7160%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7161
7162%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7163%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7164 <i>; yields {i32}:result2 = 8</i>
7165%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7166
7167%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7168</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7169
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7170</div>
7171
7172<!-- _______________________________________________________________________ -->
7173<div class="doc_subsubsection">
7174 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
7175</div>
7176<div class="doc_text">
7177<h5>Syntax:</h5>
7178
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7179<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7180 integer bit width. Not all targets support all bit widths however.</p>
7181
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7182<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7183 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7184 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7185 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7186 declare i64 @llvm.atomic.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;val&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7187</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7188
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7189<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7190<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7191 the value from memory. It then stores the value in <tt>val</tt> in the memory
7192 at <tt>ptr</tt>.</p>
7193
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7194<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7195<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7196 the <tt>val</tt> argument and the result must be integers of the same bit
7197 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7198 integer type. The targets may only lower integer representations they
7199 support.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7200
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7201<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7202<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7203 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7204 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7205
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7206<h5>Examples:</h5>
7207<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7208%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7209%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7210 store i32 4, %ptr
7211
7212%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7213%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7214 <i>; yields {i32}:result1 = 4</i>
7215%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7216%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7217
7218%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7219%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7220 <i>; yields {i32}:result2 = 8</i>
7221
7222%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7223%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7224</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7225
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7226</div>
7227
7228<!-- _______________________________________________________________________ -->
7229<div class="doc_subsubsection">
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7230 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7231
7232</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7233
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7234<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7235
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7236<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7237<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7238 any integer bit width. Not all targets support all bit widths however.</p>
7239
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7240<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7241 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7242 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7243 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7244 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7245</pre>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7246
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7247<h5>Overview:</h5>
7248<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7249 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7250
7251<h5>Arguments:</h5>
7252<p>The intrinsic takes two arguments, the first a pointer to an integer value
7253 and the second an integer value. The result is also an integer value. These
7254 integer types can have any bit width, but they must all have the same bit
7255 width. The targets may only lower integer representations they support.</p>
7256
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7257<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7258<p>This intrinsic does a series of operations atomically. It first loads the
7259 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7260 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7261
7262<h5>Examples:</h5>
7263<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7264%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7265%ptr = bitcast i8* %mallocP to i32*
7266 store i32 4, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7267%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7268 <i>; yields {i32}:result1 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7269%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7270 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7271%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7272 <i>; yields {i32}:result3 = 10</i>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7273%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7274</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7275
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7276</div>
7277
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7278<!-- _______________________________________________________________________ -->
7279<div class="doc_subsubsection">
7280 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
7281
7282</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7283
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7284<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7285
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7286<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7287<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7288 any integer bit width and for different address spaces. Not all targets
7289 support all bit widths however.</p>
7290
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7291<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7292 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7293 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7294 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7295 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7296</pre>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7297
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7298<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7299<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7300 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7301
7302<h5>Arguments:</h5>
7303<p>The intrinsic takes two arguments, the first a pointer to an integer value
7304 and the second an integer value. The result is also an integer value. These
7305 integer types can have any bit width, but they must all have the same bit
7306 width. The targets may only lower integer representations they support.</p>
7307
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7308<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7309<p>This intrinsic does a series of operations atomically. It first loads the
7310 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7311 result to <tt>ptr</tt>. It yields the original value stored
7312 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7313
7314<h5>Examples:</h5>
7315<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7316%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7317%ptr = bitcast i8* %mallocP to i32*
7318 store i32 8, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7319%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7320 <i>; yields {i32}:result1 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7321%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7322 <i>; yields {i32}:result2 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7323%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7324 <i>; yields {i32}:result3 = 2</i>
7325%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7326</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7327
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7328</div>
7329
7330<!-- _______________________________________________________________________ -->
7331<div class="doc_subsubsection">
7332 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
7333 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
7334 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
7335 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7336</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7337
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7338<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7339
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7340<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7341<p>These are overloaded intrinsics. You can
7342 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7343 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7344 bit width and for different address spaces. Not all targets support all bit
7345 widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7346
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7347<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7348 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7349 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7350 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7351 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7352</pre>
7353
7354<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7355 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7356 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7357 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7358 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7359</pre>
7360
7361<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7362 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7363 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7364 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7365 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7366</pre>
7367
7368<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7369 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7370 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7371 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7372 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7373</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7374
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7375<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7376<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7377 the value stored in memory at <tt>ptr</tt>. It yields the original value
7378 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7379
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7380<h5>Arguments:</h5>
7381<p>These intrinsics take two arguments, the first a pointer to an integer value
7382 and the second an integer value. The result is also an integer value. These
7383 integer types can have any bit width, but they must all have the same bit
7384 width. The targets may only lower integer representations they support.</p>
7385
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7386<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7387<p>These intrinsics does a series of operations atomically. They first load the
7388 value stored at <tt>ptr</tt>. They then do the bitwise
7389 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7390 original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7391
7392<h5>Examples:</h5>
7393<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7394%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7395%ptr = bitcast i8* %mallocP to i32*
7396 store i32 0x0F0F, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7397%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7398 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7399%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7400 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7401%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7402 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7403%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7404 <i>; yields {i32}:result3 = FF</i>
7405%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7406</pre>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7407
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7408</div>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7409
7410<!-- _______________________________________________________________________ -->
7411<div class="doc_subsubsection">
7412 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
7413 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
7414 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
7415 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7416</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7417
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7418<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7419
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7420<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7421<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7422 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7423 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7424 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7425
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7426<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7427 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7428 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7429 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7430 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7431</pre>
7432
7433<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7434 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7435 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7436 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7437 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7438</pre>
7439
7440<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7441 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7442 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7443 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7444 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7445</pre>
7446
7447<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7448 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7449 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7450 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7451 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7452</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7453
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7454<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7455<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7456 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7457 original value at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7458
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7459<h5>Arguments:</h5>
7460<p>These intrinsics take two arguments, the first a pointer to an integer value
7461 and the second an integer value. The result is also an integer value. These
7462 integer types can have any bit width, but they must all have the same bit
7463 width. The targets may only lower integer representations they support.</p>
7464
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7465<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7466<p>These intrinsics does a series of operations atomically. They first load the
7467 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7468 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7469 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7470
7471<h5>Examples:</h5>
7472<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7473%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7474%ptr = bitcast i8* %mallocP to i32*
7475 store i32 7, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7476%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7477 <i>; yields {i32}:result0 = 7</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7478%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7479 <i>; yields {i32}:result1 = -2</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7480%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7481 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7482%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7483 <i>; yields {i32}:result3 = 8</i>
7484%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7485</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7486
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7487</div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7488
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7489
7490<!-- ======================================================================= -->
7491<div class="doc_subsection">
7492 <a name="int_memorymarkers">Memory Use Markers</a>
7493</div>
7494
7495<div class="doc_text">
7496
7497<p>This class of intrinsics exists to information about the lifetime of memory
7498 objects and ranges where variables are immutable.</p>
7499
7500</div>
7501
7502<!-- _______________________________________________________________________ -->
7503<div class="doc_subsubsection">
7504 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
7505</div>
7506
7507<div class="doc_text">
7508
7509<h5>Syntax:</h5>
7510<pre>
7511 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7512</pre>
7513
7514<h5>Overview:</h5>
7515<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7516 object's lifetime.</p>
7517
7518<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7519<p>The first argument is a constant integer representing the size of the
7520 object, or -1 if it is variable sized. The second argument is a pointer to
7521 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7522
7523<h5>Semantics:</h5>
7524<p>This intrinsic indicates that before this point in the code, the value of the
7525 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
Nick Lewycky8d336592009-10-27 16:56:58 +0000[diff] [blame]7526 never be used and has an undefined value. A load from the pointer that
7527 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7528 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7529
7530</div>
7531
7532<!-- _______________________________________________________________________ -->
7533<div class="doc_subsubsection">
7534 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
7535</div>
7536
7537<div class="doc_text">
7538
7539<h5>Syntax:</h5>
7540<pre>
7541 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7542</pre>
7543
7544<h5>Overview:</h5>
7545<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7546 object's lifetime.</p>
7547
7548<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7549<p>The first argument is a constant integer representing the size of the
7550 object, or -1 if it is variable sized. The second argument is a pointer to
7551 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7552
7553<h5>Semantics:</h5>
7554<p>This intrinsic indicates that after this point in the code, the value of the
7555 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7556 never be used and has an undefined value. Any stores into the memory object
7557 following this intrinsic may be removed as dead.
7558
7559</div>
7560
7561<!-- _______________________________________________________________________ -->
7562<div class="doc_subsubsection">
7563 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
7564</div>
7565
7566<div class="doc_text">
7567
7568<h5>Syntax:</h5>
7569<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +0000[diff] [blame]7570 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7571</pre>
7572
7573<h5>Overview:</h5>
7574<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7575 a memory object will not change.</p>
7576
7577<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7578<p>The first argument is a constant integer representing the size of the
7579 object, or -1 if it is variable sized. The second argument is a pointer to
7580 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7581
7582<h5>Semantics:</h5>
7583<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7584 the return value, the referenced memory location is constant and
7585 unchanging.</p>
7586
7587</div>
7588
7589<!-- _______________________________________________________________________ -->
7590<div class="doc_subsubsection">
7591 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
7592</div>
7593
7594<div class="doc_text">
7595
7596<h5>Syntax:</h5>
7597<pre>
7598 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7599</pre>
7600
7601<h5>Overview:</h5>
7602<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7603 a memory object are mutable.</p>
7604
7605<h5>Arguments:</h5>
7606<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7607 The second argument is a constant integer representing the size of the
7608 object, or -1 if it is variable sized and the third argument is a pointer
7609 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7610
7611<h5>Semantics:</h5>
7612<p>This intrinsic indicates that the memory is mutable again.</p>
7613
7614</div>
7615
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7616<!-- ======================================================================= -->
7617<div class="doc_subsection">
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7618 <a name="int_general">General Intrinsics</a>
7619</div>
7620
7621<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7622
7623<p>This class of intrinsics is designed to be generic and has no specific
7624 purpose.</p>
7625
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7626</div>
7627
7628<!-- _______________________________________________________________________ -->
7629<div class="doc_subsubsection">
7630 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
7631</div>
7632
7633<div class="doc_text">
7634
7635<h5>Syntax:</h5>
7636<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7637 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 +0000[diff] [blame]7638</pre>
7639
7640<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7641<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7642
7643<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7644<p>The first argument is a pointer to a value, the second is a pointer to a
7645 global string, the third is a pointer to a global string which is the source
7646 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7647
7648<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7649<p>This intrinsic allows annotation of local variables with arbitrary strings.
7650 This can be useful for special purpose optimizations that want to look for
7651 these annotations. These have no other defined use, they are ignored by code
7652 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7653
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7654</div>
7655
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]7656<!-- _______________________________________________________________________ -->
7657<div class="doc_subsubsection">
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]7658 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]7659</div>
7660
7661<div class="doc_text">
7662
7663<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7664<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
7665 any integer bit width.</p>
7666
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]7667<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7668 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7669 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7670 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7671 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
7672 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 +0000[diff] [blame]7673</pre>
7674
7675<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7676<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]7677
7678<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7679<p>The first argument is an integer value (result of some expression), the
7680 second is a pointer to a global string, the third is a pointer to a global
7681 string which is the source file name, and the last argument is the line
7682 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]7683
7684<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7685<p>This intrinsic allows annotations to be put on arbitrary expressions with
7686 arbitrary strings. This can be useful for special purpose optimizations that
7687 want to look for these annotations. These have no other defined use, they
7688 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]7689
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]7690</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7691
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]7692<!-- _______________________________________________________________________ -->
7693<div class="doc_subsubsection">
7694 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
7695</div>
7696
7697<div class="doc_text">
7698
7699<h5>Syntax:</h5>
7700<pre>
7701 declare void @llvm.trap()
7702</pre>
7703
7704<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7705<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]7706
7707<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7708<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]7709
7710<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7711<p>This intrinsics is lowered to the target dependent trap instruction. If the
7712 target does not have a trap instruction, this intrinsic will be lowered to
7713 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]7714
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]7715</div>
7716
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7717<!-- _______________________________________________________________________ -->
7718<div class="doc_subsubsection">
Misha Brukmandccb0252008-11-22 23:55:29 +0000[diff] [blame]7719 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7720</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7721
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7722<div class="doc_text">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7723
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7724<h5>Syntax:</h5>
7725<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7726 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7727</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7728
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7729<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7730<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
7731 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
7732 ensure that it is placed on the stack before local variables.</p>
7733
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7734<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7735<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
7736 arguments. The first argument is the value loaded from the stack
7737 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
7738 that has enough space to hold the value of the guard.</p>
7739
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7740<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7741<p>This intrinsic causes the prologue/epilogue inserter to force the position of
7742 the <tt>AllocaInst</tt> stack slot to be before local variables on the
7743 stack. This is to ensure that if a local variable on the stack is
7744 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]7745 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7746 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
7747 function.</p>
7748
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]7749</div>
7750
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]7751<!-- _______________________________________________________________________ -->
7752<div class="doc_subsubsection">
7753 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
7754</div>
7755
7756<div class="doc_text">
7757
7758<h5>Syntax:</h5>
7759<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7760 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
7761 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]7762</pre>
7763
7764<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]7765<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
7766 the optimizers to determine at compile time whether a) an operation (like
7767 memcpy) will overflow a buffer that corresponds to an object, or b) that a
7768 runtime check for overflow isn't necessary. An object in this context means
7769 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]7770
7771<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]7772<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]7773 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]7774 is a boolean 0 or 1. This argument determines whether you want the
7775 maximum (0) or minimum (1) bytes remaining. This needs to be a literal 0 or
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]7776 1, variables are not allowed.</p>
7777
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]7778<h5>Semantics:</h5>
7779<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]7780 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
7781 depending on the <tt>type</tt> argument, if the size cannot be determined at
7782 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]7783
7784</div>
7785
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]7786<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]7787<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]7788<address>
7789 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]7790 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]7791 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]7792 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]7793
7794 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]7795 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]7796 Last modified: $Date$
7797</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]7798
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]7799</body>
7800</html>