Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 25456ef74ca7ac9bf46451f28d995f785e2596de / . / docs / LangRef.html
blob: 7882da67e1aa93b12ec05ddf3262e76580b9ba9d [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
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]15<h1>LLVM Language Reference Manual</h1>
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>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]56 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]57 <li><a href="#ordering">Atomic Memory Ordering Constraints</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]58 </ol>
59 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]60 <li><a href="#typesystem">Type System</a>
61 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]62 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]63 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]64 <ol>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]65 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]66 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]67 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]68 <li><a href="#t_void">Void Type</a></li>
69 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]70 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]71 </ol>
72 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]73 <li><a href="#t_derived">Derived Types</a>
74 <ol>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]75 <li><a href="#t_aggregate">Aggregate Types</a>
76 <ol>
77 <li><a href="#t_array">Array Type</a></li>
78 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]79 <li><a href="#t_opaque">Opaque Structure Types</a></li>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]80 <li><a href="#t_vector">Vector Type</a></li>
81 </ol>
82 </li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]83 <li><a href="#t_function">Function Type</a></li>
84 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]85 </ol>
86 </li>
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>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]127 <li><a href="#i_resume">'<tt>resume</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]128 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]129 </ol>
130 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]131 <li><a href="#binaryops">Binary Operations</a>
132 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]133 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]134 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]135 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]136 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]137 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]138 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]139 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
140 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
141 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]142 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
143 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
144 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]145 </ol>
146 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]147 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
148 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]149 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
150 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
151 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]152 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]153 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]154 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]155 </ol>
156 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]157 <li><a href="#vectorops">Vector Operations</a>
158 <ol>
159 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
160 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
161 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]162 </ol>
163 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]164 <li><a href="#aggregateops">Aggregate Operations</a>
165 <ol>
166 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
167 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
168 </ol>
169 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]170 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]171 <ol>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]172 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
173 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
174 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
175 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
176 <li><a href="#i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a></li>
177 <li><a href="#i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]178 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]179 </ol>
180 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]181 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]182 <ol>
183 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
184 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
185 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
186 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
187 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]188 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
189 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
190 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
191 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]192 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
193 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]194 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]195 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]196 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]197 <li><a href="#otherops">Other Operations</a>
198 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]199 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
200 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]201 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]202 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]203 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]204 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]205 <li><a href="#i_landingpad">'<tt>landingpad</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]206 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]207 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]208 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]209 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]210 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]211 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]212 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
213 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]214 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
215 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
216 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]217 </ol>
218 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]219 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
220 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]221 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
222 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
223 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]224 </ol>
225 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]226 <li><a href="#int_codegen">Code Generator Intrinsics</a>
227 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]228 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
229 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
230 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
231 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
232 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
233 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000[diff] [blame]234 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]235 </ol>
236 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]237 <li><a href="#int_libc">Standard C Library Intrinsics</a>
238 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]239 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
242 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
243 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]244 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
245 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
246 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohman08b280b2011-05-27 00:36:31 +0000[diff] [blame]247 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
248 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]249 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]250 </ol>
251 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]252 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]253 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]254 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]255 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
256 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
257 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]258 </ol>
259 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]260 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
261 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]262 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
263 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
264 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
265 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
266 <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]267 <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]268 </ol>
269 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]270 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
271 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]272 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
273 <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]274 </ol>
275 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]276 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]277 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]278 <li><a href="#int_trampoline">Trampoline Intrinsics</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]279 <ol>
280 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]281 <li><a href="#int_at">'<tt>llvm.adjust.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]282 </ol>
283 </li>
Bill Wendling3c44f5b2008-11-18 22:10:53 +0000[diff] [blame]284 <li><a href="#int_atomics">Atomic intrinsics</a>
285 <ol>
286 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
287 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
288 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
289 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
290 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
291 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
292 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
293 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
294 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
295 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
296 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
297 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
298 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
299 </ol>
300 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]301 <li><a href="#int_memorymarkers">Memory Use Markers</a>
302 <ol>
303 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
304 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
305 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
306 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
307 </ol>
308 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]309 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]310 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]311 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]312 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]313 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]314 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]315 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]316 '<tt>llvm.trap</tt>' Intrinsic</a></li>
317 <li><a href="#int_stackprotector">
318 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]319 <li><a href="#int_objectsize">
320 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]321 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]322 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]323 </ol>
324 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]325</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]326
327<div class="doc_author">
328 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
329 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]330</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]331
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]332<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]333<h2><a name="abstract">Abstract</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]334<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]335
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]336<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]337
338<p>This document is a reference manual for the LLVM assembly language. LLVM is
339 a Static Single Assignment (SSA) based representation that provides type
340 safety, low-level operations, flexibility, and the capability of representing
341 'all' high-level languages cleanly. It is the common code representation
342 used throughout all phases of the LLVM compilation strategy.</p>
343
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]344</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]345
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]346<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]347<h2><a name="introduction">Introduction</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]348<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]349
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]350<div>
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 code representation is designed to be used in three different forms:
353 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
354 for fast loading by a Just-In-Time compiler), and as a human readable
355 assembly language representation. This allows LLVM to provide a powerful
356 intermediate representation for efficient compiler transformations and
357 analysis, while providing a natural means to debug and visualize the
358 transformations. The three different forms of LLVM are all equivalent. This
359 document describes the human readable representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]360
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]361<p>The LLVM representation aims to be light-weight and low-level while being
362 expressive, typed, and extensible at the same time. It aims to be a
363 "universal IR" of sorts, by being at a low enough level that high-level ideas
364 may be cleanly mapped to it (similar to how microprocessors are "universal
365 IR's", allowing many source languages to be mapped to them). By providing
366 type information, LLVM can be used as the target of optimizations: for
367 example, through pointer analysis, it can be proven that a C automatic
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]368 variable is never accessed outside of the current function, allowing it to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]369 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]370
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]371<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]372<h4>
373 <a name="wellformed">Well-Formedness</a>
374</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]375
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]376<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]377
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]378<p>It is important to note that this document describes 'well formed' LLVM
379 assembly language. There is a difference between what the parser accepts and
380 what is considered 'well formed'. For example, the following instruction is
381 syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]382
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]383<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]384%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]385</pre>
386
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]387<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
388 LLVM infrastructure provides a verification pass that may be used to verify
389 that an LLVM module is well formed. This pass is automatically run by the
390 parser after parsing input assembly and by the optimizer before it outputs
391 bitcode. The violations pointed out by the verifier pass indicate bugs in
392 transformation passes or input to the parser.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]393
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]394</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]395
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]396</div>
397
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame]398<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]399
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]400<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]401<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]402<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]403
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]404<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]405
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]406<p>LLVM identifiers come in two basic types: global and local. Global
407 identifiers (functions, global variables) begin with the <tt>'@'</tt>
408 character. Local identifiers (register names, types) begin with
409 the <tt>'%'</tt> character. Additionally, there are three different formats
410 for identifiers, for different purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]411
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]412<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]413 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]414 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
415 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
416 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
417 other characters in their names can be surrounded with quotes. Special
418 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
419 ASCII code for the character in hexadecimal. In this way, any character
420 can be used in a name value, even quotes themselves.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]421
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]422 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]423 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]424
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]425 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]426 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]427</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]428
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]429<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]430 don't need to worry about name clashes with reserved words, and the set of
431 reserved words may be expanded in the future without penalty. Additionally,
432 unnamed identifiers allow a compiler to quickly come up with a temporary
433 variable without having to avoid symbol table conflicts.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]434
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]435<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]436 languages. There are keywords for different opcodes
437 ('<tt><a href="#i_add">add</a></tt>',
438 '<tt><a href="#i_bitcast">bitcast</a></tt>',
439 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
440 ('<tt><a href="#t_void">void</a></tt>',
441 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
442 reserved words cannot conflict with variable names, because none of them
443 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]444
445<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]446 '<tt>%X</tt>' by 8:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]447
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]448<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]449
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]450<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]451%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]452</pre>
453
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]454<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]455
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]456<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]457%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]458</pre>
459
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]460<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]461
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]462<pre class="doc_code">
Gabor Greifec58f752009-10-28 13:05:07 +0000[diff] [blame]463%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
464%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]465%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]466</pre>
467
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]468<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
469 lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]470
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]471<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]472 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]473 line.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]474
475 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]476 assigned to a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]477
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]478 <li>Unnamed temporaries are numbered sequentially</li>
479</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]480
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]481<p>It also shows a convention that we follow in this document. When
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]482 demonstrating instructions, we will follow an instruction with a comment that
483 defines the type and name of value produced. Comments are shown in italic
484 text.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]485
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]486</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]487
488<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]489<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]490<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]491<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]492<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]493<h3>
494 <a name="modulestructure">Module Structure</a>
495</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]496
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]497<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]498
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]499<p>LLVM programs are composed of "Module"s, each of which is a translation unit
500 of the input programs. Each module consists of functions, global variables,
501 and symbol table entries. Modules may be combined together with the LLVM
502 linker, which merges function (and global variable) definitions, resolves
503 forward declarations, and merges symbol table entries. Here is an example of
504 the "hello world" module:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]505
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]506<pre class="doc_code">
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]507<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckydb9cd762011-01-29 01:09:53 +0000[diff] [blame]508<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]509
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]510<i>; External declaration of the puts function</i>&nbsp;
511<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]512
513<i>; Definition of main function</i>
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]514define i32 @main() { <i>; i32()* </i>&nbsp;
515 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
516 %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]517
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]518 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
519 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
520 <a href="#i_ret">ret</a> i32 0&nbsp;
521}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]522
523<i>; Named metadata</i>
524!1 = metadata !{i32 41}
525!foo = !{!1, null}
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]526</pre>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]527
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]528<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]529 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]530 a <a href="#functionstructure">function definition</a> for
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]531 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
532 "<tt>foo"</tt>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]533
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]534<p>In general, a module is made up of a list of global values, where both
535 functions and global variables are global values. Global values are
536 represented by a pointer to a memory location (in this case, a pointer to an
537 array of char, and a pointer to a function), and have one of the
538 following <a href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]539
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]540</div>
541
542<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]543<h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]544 <a name="linkage">Linkage Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]545</h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]546
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]547<div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]548
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]549<p>All Global Variables and Functions have one of the following types of
550 linkage:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]551
552<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]553 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]554 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
555 by objects in the current module. In particular, linking code into a
556 module with an private global value may cause the private to be renamed as
557 necessary to avoid collisions. Because the symbol is private to the
558 module, all references can be updated. This doesn't show up in any symbol
559 table in the object file.</dd>
Rafael Espindolabb46f522009-01-15 20:18:42 +0000[diff] [blame]560
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]561 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]562 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
563 assembler and evaluated by the linker. Unlike normal strong symbols, they
564 are removed by the linker from the final linked image (executable or
565 dynamic library).</dd>
566
567 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
568 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
569 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
570 linker. The symbols are removed by the linker from the final linked image
571 (executable or dynamic library).</dd>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]572
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]573 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
574 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
575 of the object is not taken. For instance, functions that had an inline
576 definition, but the compiler decided not to inline it. Note,
577 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
578 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
579 visibility. The symbols are removed by the linker from the final linked
580 image (executable or dynamic library).</dd>
581
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]582 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling07d31772010-06-29 22:34:52 +0000[diff] [blame]583 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]584 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
585 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]586
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]587 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]588 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]589 into the object file corresponding to the LLVM module. They exist to
590 allow inlining and other optimizations to take place given knowledge of
591 the definition of the global, which is known to be somewhere outside the
592 module. Globals with <tt>available_externally</tt> linkage are allowed to
593 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
594 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]595
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]596 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]597 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner873187c2010-01-09 19:15:14 +0000[diff] [blame]598 the same name when linkage occurs. This can be used to implement
599 some forms of inline functions, templates, or other code which must be
600 generated in each translation unit that uses it, but where the body may
601 be overridden with a more definitive definition later. Unreferenced
602 <tt>linkonce</tt> globals are allowed to be discarded. Note that
603 <tt>linkonce</tt> linkage does not actually allow the optimizer to
604 inline the body of this function into callers because it doesn't know if
605 this definition of the function is the definitive definition within the
606 program or whether it will be overridden by a stronger definition.
607 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
608 linkage.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]609
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]610 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]611 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
612 <tt>linkonce</tt> linkage, except that unreferenced globals with
613 <tt>weak</tt> linkage may not be discarded. This is used for globals that
614 are declared "weak" in C source code.</dd>
615
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]616 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]617 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
618 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
619 global scope.
620 Symbols with "<tt>common</tt>" linkage are merged in the same way as
621 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]622 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]623 must have a zero initializer, and may not be marked '<a
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]624 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
625 have common linkage.</dd>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]626
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]627
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]628 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]629 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]630 pointer to array type. When two global variables with appending linkage
631 are linked together, the two global arrays are appended together. This is
632 the LLVM, typesafe, equivalent of having the system linker append together
633 "sections" with identical names when .o files are linked.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]634
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]635 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]636 <dd>The semantics of this linkage follow the ELF object file model: the symbol
637 is weak until linked, if not linked, the symbol becomes null instead of
638 being an undefined reference.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]639
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]640 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
641 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]642 <dd>Some languages allow differing globals to be merged, such as two functions
643 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]644 that only equivalent globals are ever merged (the "one definition rule"
645 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]646 and <tt>weak_odr</tt> linkage types to indicate that the global will only
647 be merged with equivalent globals. These linkage types are otherwise the
648 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]649
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]650 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]651 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]652 visible, meaning that it participates in linkage and can be used to
653 resolve external symbol references.</dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]654</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]655
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]656<p>The next two types of linkage are targeted for Microsoft Windows platform
657 only. They are designed to support importing (exporting) symbols from (to)
658 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]659
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]660<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]661 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]662 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]663 or variable via a global pointer to a pointer that is set up by the DLL
664 exporting the symbol. On Microsoft Windows targets, the pointer name is
665 formed by combining <code>__imp_</code> and the function or variable
666 name.</dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]667
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]668 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]669 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]670 pointer to a pointer in a DLL, so that it can be referenced with the
671 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
672 name is formed by combining <code>__imp_</code> and the function or
673 variable name.</dd>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]674</dl>
675
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]676<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
677 another module defined a "<tt>.LC0</tt>" variable and was linked with this
678 one, one of the two would be renamed, preventing a collision. Since
679 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
680 declarations), they are accessible outside of the current module.</p>
681
682<p>It is illegal for a function <i>declaration</i> to have any linkage type
683 other than "externally visible", <tt>dllimport</tt>
684 or <tt>extern_weak</tt>.</p>
685
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]686<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]687 or <tt>weak_odr</tt> linkages.</p>
688
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]689</div>
690
691<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]692<h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]693 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]694</h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]695
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]696<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]697
698<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]699 and <a href="#i_invoke">invokes</a> can all have an optional calling
700 convention specified for the call. The calling convention of any pair of
701 dynamic caller/callee must match, or the behavior of the program is
702 undefined. The following calling conventions are supported by LLVM, and more
703 may be added in the future:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]704
705<dl>
706 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]707 <dd>This calling convention (the default if no other calling convention is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]708 specified) matches the target C calling conventions. This calling
709 convention supports varargs function calls and tolerates some mismatch in
710 the declared prototype and implemented declaration of the function (as
711 does normal C).</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]712
713 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]714 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]715 (e.g. by passing things in registers). This calling convention allows the
716 target to use whatever tricks it wants to produce fast code for the
717 target, without having to conform to an externally specified ABI
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]718 (Application Binary Interface).
719 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]720 when this or the GHC convention is used.</a> This calling convention
721 does not support varargs and requires the prototype of all callees to
722 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]723
724 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]725 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]726 as possible under the assumption that the call is not commonly executed.
727 As such, these calls often preserve all registers so that the call does
728 not break any live ranges in the caller side. This calling convention
729 does not support varargs and requires the prototype of all callees to
730 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]731
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]732 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
733 <dd>This calling convention has been implemented specifically for use by the
734 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
735 It passes everything in registers, going to extremes to achieve this by
736 disabling callee save registers. This calling convention should not be
737 used lightly but only for specific situations such as an alternative to
738 the <em>register pinning</em> performance technique often used when
739 implementing functional programming languages.At the moment only X86
740 supports this convention and it has the following limitations:
741 <ul>
742 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
743 floating point types are supported.</li>
744 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
745 6 floating point parameters.</li>
746 </ul>
747 This calling convention supports
748 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
749 requires both the caller and callee are using it.
750 </dd>
751
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]752 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]753 <dd>Any calling convention may be specified by number, allowing
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]754 target-specific calling conventions to be used. Target specific calling
755 conventions start at 64.</dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]756</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]757
758<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]759 support Pascal conventions or any other well-known target-independent
760 convention.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]761
762</div>
763
764<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]765<h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]766 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]767</h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]768
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]769<div>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]770
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]771<p>All Global Variables and Functions have one of the following visibility
772 styles:</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]773
774<dl>
775 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]776 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]777 that the declaration is visible to other modules and, in shared libraries,
778 means that the declared entity may be overridden. On Darwin, default
779 visibility means that the declaration is visible to other modules. Default
780 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]781
782 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]783 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]784 object if they are in the same shared object. Usually, hidden visibility
785 indicates that the symbol will not be placed into the dynamic symbol
786 table, so no other module (executable or shared library) can reference it
787 directly.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]788
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]789 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]790 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]791 the dynamic symbol table, but that references within the defining module
792 will bind to the local symbol. That is, the symbol cannot be overridden by
793 another module.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]794</dl>
795
796</div>
797
798<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]799<h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]800 <a name="namedtypes">Named Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]801</h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]802
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]803<div>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]804
805<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]806 it easier to read the IR and make the IR more condensed (particularly when
807 recursive types are involved). An example of a name specification is:</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]808
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]809<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]810%mytype = type { %mytype*, i32 }
811</pre>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]812
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]813<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattnerdc65f222010-08-17 23:26:04 +0000[diff] [blame]814 "<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]815 is expected with the syntax "%mytype".</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]816
817<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]818 and that you can therefore specify multiple names for the same type. This
819 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
820 uses structural typing, the name is not part of the type. When printing out
821 LLVM IR, the printer will pick <em>one name</em> to render all types of a
822 particular shape. This means that if you have code where two different
823 source types end up having the same LLVM type, that the dumper will sometimes
824 print the "wrong" or unexpected type. This is an important design point and
825 isn't going to change.</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]826
827</div>
828
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]829<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]830<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]831 <a name="globalvars">Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]832</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]833
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]834<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]835
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]836<p>Global variables define regions of memory allocated at compilation time
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]837 instead of run-time. Global variables may optionally be initialized, may
838 have an explicit section to be placed in, and may have an optional explicit
839 alignment specified. A variable may be defined as "thread_local", which
840 means that it will not be shared by threads (each thread will have a
841 separated copy of the variable). A variable may be defined as a global
842 "constant," which indicates that the contents of the variable
843 will <b>never</b> be modified (enabling better optimization, allowing the
844 global data to be placed in the read-only section of an executable, etc).
845 Note that variables that need runtime initialization cannot be marked
846 "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]847
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]848<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
849 constant, even if the final definition of the global is not. This capability
850 can be used to enable slightly better optimization of the program, but
851 requires the language definition to guarantee that optimizations based on the
852 'constantness' are valid for the translation units that do not include the
853 definition.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]854
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]855<p>As SSA values, global variables define pointer values that are in scope
856 (i.e. they dominate) all basic blocks in the program. Global variables
857 always define a pointer to their "content" type because they describe a
858 region of memory, and all memory objects in LLVM are accessed through
859 pointers.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]860
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]861<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
862 that the address is not significant, only the content. Constants marked
Rafael Espindolaa5eaa862011-01-15 08:20:57 +0000[diff] [blame]863 like this can be merged with other constants if they have the same
864 initializer. Note that a constant with significant address <em>can</em>
865 be merged with a <tt>unnamed_addr</tt> constant, the result being a
866 constant whose address is significant.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]867
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]868<p>A global variable may be declared to reside in a target-specific numbered
869 address space. For targets that support them, address spaces may affect how
870 optimizations are performed and/or what target instructions are used to
871 access the variable. The default address space is zero. The address space
872 qualifier must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000[diff] [blame]873
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]874<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]875 supports it, it will emit globals to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]876
Chris Lattnerce99fa92010-04-28 00:13:42 +0000[diff] [blame]877<p>An explicit alignment may be specified for a global, which must be a power
878 of 2. If not present, or if the alignment is set to zero, the alignment of
879 the global is set by the target to whatever it feels convenient. If an
880 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner2d4b8ee2010-04-28 00:31:12 +0000[diff] [blame]881 alignment. Targets and optimizers are not allowed to over-align the global
882 if the global has an assigned section. In this case, the extra alignment
883 could be observable: for example, code could assume that the globals are
884 densely packed in their section and try to iterate over them as an array,
885 alignment padding would break this iteration.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]886
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]887<p>For example, the following defines a global in a numbered address space with
888 an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]889
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]890<pre class="doc_code">
Dan Gohman398873c2009-01-11 00:40:00 +0000[diff] [blame]891@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]892</pre>
893
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]894</div>
895
896
897<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]898<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]899 <a name="functionstructure">Functions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]900</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]901
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]902<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]903
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]904<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]905 optional <a href="#linkage">linkage type</a>, an optional
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) argument list (each with optional
911 <a href="#paramattrs">parameter attributes</a>), optional
912 <a href="#fnattrs">function attributes</a>, an optional section, an optional
913 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
914 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]915
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]916<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
917 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]918 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]919 <a href="#callingconv">calling convention</a>,
920 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]921 <a href="#paramattrs">parameter attribute</a> for the return type, a function
922 name, a possibly empty list of arguments, an optional alignment, and an
923 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]924
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]925<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]926 (Control Flow Graph) for the function. Each basic block may optionally start
927 with a label (giving the basic block a symbol table entry), contains a list
928 of instructions, and ends with a <a href="#terminators">terminator</a>
929 instruction (such as a branch or function return).</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]930
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]931<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]932 executed on entrance to the function, and it is not allowed to have
933 predecessor basic blocks (i.e. there can not be any branches to the entry
934 block of a function). Because the block can have no predecessors, it also
935 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]936
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]937<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]938 supports it, it will emit functions to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]939
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]940<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]941 the alignment is set to zero, the alignment of the function is set by the
942 target to whatever it feels convenient. If an explicit alignment is
943 specified, the function is forced to have at least that much alignment. All
944 alignments must be a power of 2.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]945
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]946<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
947 be significant and two identical functions can be merged</p>.
948
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]949<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]950<pre class="doc_code">
Chris Lattner50ad45c2008-10-13 16:55:18 +0000[diff] [blame]951define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]952 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
953 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
954 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
955 [<a href="#gc">gc</a>] { ... }
956</pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]957
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]958</div>
959
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]960<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]961<h3>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]962 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]963</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]964
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]965<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]966
967<p>Aliases act as "second name" for the aliasee value (which can be either
968 function, global variable, another alias or bitcast of global value). Aliases
969 may have an optional <a href="#linkage">linkage type</a>, and an
970 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]971
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]972<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]973<pre class="doc_code">
Duncan Sands0b23ac12008-09-12 20:48:21 +0000[diff] [blame]974@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]975</pre>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]976
977</div>
978
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]979<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]980<h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]981 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]982</h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]983
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]984<div>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]985
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]986<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]987 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]988 a named metadata.</p>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]989
990<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]991<pre class="doc_code">
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]992; Some unnamed metadata nodes, which are referenced by the named metadata.
993!0 = metadata !{metadata !"zero"}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]994!1 = metadata !{metadata !"one"}
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]995!2 = metadata !{metadata !"two"}
Dan Gohman1005bc52010-07-13 19:48:13 +0000[diff] [blame]996; A named metadata.
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]997!name = !{!0, !1, !2}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]998</pre>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]999
1000</div>
1001
1002<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1003<h3>
1004 <a name="paramattrs">Parameter Attributes</a>
1005</h3>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1006
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1007<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1008
1009<p>The return type and each parameter of a function type may have a set of
1010 <i>parameter attributes</i> associated with them. Parameter attributes are
1011 used to communicate additional information about the result or parameters of
1012 a function. Parameter attributes are considered to be part of the function,
1013 not of the function type, so functions with different parameter attributes
1014 can have the same function type.</p>
1015
1016<p>Parameter attributes are simple keywords that follow the type specified. If
1017 multiple parameter attributes are needed, they are space separated. For
1018 example:</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1019
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1020<pre class="doc_code">
Nick Lewyckyb6a7d252009-02-15 23:06:14 +0000[diff] [blame]1021declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner66d922c2008-10-04 18:33:34 +0000[diff] [blame]1022declare i32 @atoi(i8 zeroext)
1023declare signext i8 @returns_signed_char()
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1024</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1025
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1026<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1027 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1028
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1029<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1030
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1031<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1032 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1033 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichebe81732011-03-16 22:20:18 +0000[diff] [blame]1034 should be zero-extended to the extent required by the target's ABI (which
1035 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1036 parameter) or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1037
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1038 <dt><tt><b>signext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1039 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich9e69ff92011-03-17 14:21:58 +0000[diff] [blame]1040 should be sign-extended to the extent required by the target's ABI (which
1041 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1042 return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1043
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1044 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1045 <dd>This indicates that this parameter or return value should be treated in a
1046 special target-dependent fashion during while emitting code for a function
1047 call or return (usually, by putting it in a register as opposed to memory,
1048 though some targets use it to distinguish between two different kinds of
1049 registers). Use of this attribute is target-specific.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1050
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1051 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1052 <dd><p>This indicates that the pointer parameter should really be passed by
1053 value to the function. The attribute implies that a hidden copy of the
1054 pointee
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1055 is made between the caller and the callee, so the callee is unable to
1056 modify the value in the callee. This attribute is only valid on LLVM
1057 pointer arguments. It is generally used to pass structs and arrays by
1058 value, but is also valid on pointers to scalars. The copy is considered
1059 to belong to the caller not the callee (for example,
1060 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1061 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1062 values.</p>
1063
1064 <p>The byval attribute also supports specifying an alignment with
1065 the align attribute. It indicates the alignment of the stack slot to
1066 form and the known alignment of the pointer specified to the call site. If
1067 the alignment is not specified, then the code generator makes a
1068 target-specific assumption.</p></dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1069
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1070 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1071 <dd>This indicates that the pointer parameter specifies the address of a
1072 structure that is the return value of the function in the source program.
1073 This pointer must be guaranteed by the caller to be valid: loads and
1074 stores to the structure may be assumed by the callee to not to trap. This
1075 may only be applied to the first parameter. This is not a valid attribute
1076 for return values. </dd>
1077
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1078 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1079 <dd>This indicates that pointer values
1080 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmanefca7f92010-07-02 23:46:54 +0000[diff] [blame]1081 value do not alias pointer values which are not <i>based</i> on it,
1082 ignoring certain "irrelevant" dependencies.
1083 For a call to the parent function, dependencies between memory
1084 references from before or after the call and from those during the call
1085 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1086 return value used in that call.
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1087 The caller shares the responsibility with the callee for ensuring that
1088 these requirements are met.
1089 For further details, please see the discussion of the NoAlias response in
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1090 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1091<br>
John McCall191d4ee2010-07-06 21:07:14 +0000[diff] [blame]1092 Note that this definition of <tt>noalias</tt> is intentionally
1093 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner211244a2010-07-06 20:51:35 +0000[diff] [blame]1094 arguments, though it is slightly weaker.
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1095<br>
1096 For function return values, C99's <tt>restrict</tt> is not meaningful,
1097 while LLVM's <tt>noalias</tt> is.
1098 </dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1099
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1100 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1101 <dd>This indicates that the callee does not make any copies of the pointer
1102 that outlive the callee itself. This is not a valid attribute for return
1103 values.</dd>
1104
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1105 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1106 <dd>This indicates that the pointer parameter can be excised using the
1107 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1108 attribute for return values.</dd>
1109</dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1110
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1111</div>
1112
1113<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1114<h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1115 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1116</h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1117
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1118<div>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1119
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1120<p>Each function may specify a garbage collector name, which is simply a
1121 string:</p>
1122
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1123<pre class="doc_code">
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1124define void @f() gc "name" { ... }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1125</pre>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1126
1127<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1128 collector which will cause the compiler to alter its output in order to
1129 support the named garbage collection algorithm.</p>
1130
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1131</div>
1132
1133<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1134<h3>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1135 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1136</h3>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1137
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1138<div>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1139
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1140<p>Function attributes are set to communicate additional information about a
1141 function. Function attributes are considered to be part of the function, not
1142 of the function type, so functions with different parameter attributes can
1143 have the same function type.</p>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1144
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1145<p>Function attributes are simple keywords that follow the type specified. If
1146 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1147
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1148<pre class="doc_code">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1149define void @f() noinline { ... }
1150define void @f() alwaysinline { ... }
1151define void @f() alwaysinline optsize { ... }
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1152define void @f() optsize { ... }
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1153</pre>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1154
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1155<dl>
Charles Davis1e063d12010-02-12 00:31:15 +0000[diff] [blame]1156 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1157 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1158 the backend should forcibly align the stack pointer. Specify the
1159 desired alignment, which must be a power of two, in parentheses.
1160
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1161 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1162 <dd>This attribute indicates that the inliner should attempt to inline this
1163 function into callers whenever possible, ignoring any active inlining size
1164 threshold for this caller.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1165
Charles Davis970bfcc2010-10-25 15:37:09 +0000[diff] [blame]1166 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis6f12e292010-10-25 16:29:03 +0000[diff] [blame]1167 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis0076d202010-10-25 19:07:39 +0000[diff] [blame]1168 meaning the function can be patched and/or hooked even while it is
1169 loaded into memory. On x86, the function prologue will be preceded
1170 by six bytes of padding and will begin with a two-byte instruction.
1171 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1172 higher were compiled in this fashion.</dd>
Charles Davis970bfcc2010-10-25 15:37:09 +0000[diff] [blame]1173
Dan Gohman129bd562011-06-16 16:03:13 +0000[diff] [blame]1174 <dt><tt><b>nonlazybind</b></tt></dt>
1175 <dd>This attribute suppresses lazy symbol binding for the function. This
1176 may make calls to the function faster, at the cost of extra program
1177 startup time if the function is not called during program startup.</dd>
1178
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +0000[diff] [blame]1179 <dt><tt><b>inlinehint</b></tt></dt>
1180 <dd>This attribute indicates that the source code contained a hint that inlining
1181 this function is desirable (such as the "inline" keyword in C/C++). It
1182 is just a hint; it imposes no requirements on the inliner.</dd>
1183
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1184 <dt><tt><b>naked</b></tt></dt>
1185 <dd>This attribute disables prologue / epilogue emission for the function.
1186 This can have very system-specific consequences.</dd>
1187
1188 <dt><tt><b>noimplicitfloat</b></tt></dt>
1189 <dd>This attributes disables implicit floating point instructions.</dd>
1190
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1191 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1192 <dd>This attribute indicates that the inliner should never inline this
1193 function in any situation. This attribute may not be used together with
1194 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1195
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1196 <dt><tt><b>noredzone</b></tt></dt>
1197 <dd>This attribute indicates that the code generator should not use a red
1198 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1199
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1200 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1201 <dd>This function attribute indicates that the function never returns
1202 normally. This produces undefined behavior at runtime if the function
1203 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +0000[diff] [blame]1204
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1205 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1206 <dd>This function attribute indicates that the function never returns with an
1207 unwind or exceptional control flow. If the function does unwind, its
1208 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +0000[diff] [blame]1209
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1210 <dt><tt><b>optsize</b></tt></dt>
1211 <dd>This attribute suggests that optimization passes and code generator passes
1212 make choices that keep the code size of this function low, and otherwise
1213 do optimizations specifically to reduce code size.</dd>
1214
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1215 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1216 <dd>This attribute indicates that the function computes its result (or decides
1217 to unwind an exception) based strictly on its arguments, without
1218 dereferencing any pointer arguments or otherwise accessing any mutable
1219 state (e.g. memory, control registers, etc) visible to caller functions.
1220 It does not write through any pointer arguments
1221 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1222 changes any state visible to callers. This means that it cannot unwind
1223 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1224 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +0000[diff] [blame]1225
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1226 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1227 <dd>This attribute indicates that the function does not write through any
1228 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1229 arguments) or otherwise modify any state (e.g. memory, control registers,
1230 etc) visible to caller functions. It may dereference pointer arguments
1231 and read state that may be set in the caller. A readonly function always
1232 returns the same value (or unwinds an exception identically) when called
1233 with the same set of arguments and global state. It cannot unwind an
1234 exception by calling the <tt>C++</tt> exception throwing methods, but may
1235 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +0000[diff] [blame]1236
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1237 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1238 <dd>This attribute indicates that the function should emit a stack smashing
1239 protector. It is in the form of a "canary"&mdash;a random value placed on
1240 the stack before the local variables that's checked upon return from the
1241 function to see if it has been overwritten. A heuristic is used to
1242 determine if a function needs stack protectors or not.<br>
1243<br>
1244 If a function that has an <tt>ssp</tt> attribute is inlined into a
1245 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1246 function will have an <tt>ssp</tt> attribute.</dd>
1247
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1248 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1249 <dd>This attribute indicates that the function should <em>always</em> emit a
1250 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1251 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1252<br>
1253 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1254 function that doesn't have an <tt>sspreq</tt> attribute or which has
1255 an <tt>ssp</tt> attribute, then the resulting function will have
1256 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindolafbff0ec2011-07-25 15:27:59 +0000[diff] [blame]1257
1258 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1259 <dd>This attribute indicates that the ABI being targeted requires that
1260 an unwind table entry be produce for this function even if we can
1261 show that no exceptions passes by it. This is normally the case for
1262 the ELF x86-64 abi, but it can be disabled for some compilation
1263 units.</dd>
1264
Rafael Espindola25456ef2011-10-03 14:45:37 +0000[diff] [blame^]1265 <dt><tt><b><a name="returns_twice">returns_twice</a></b></tt></dt>
1266 <dd>This attribute indicates that this function can return
1267 twice. The C <code>setjmp</code> is an example of such a function.
1268 The compiler disables some optimizations (like tail calls) in the caller of
1269 these functions.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1270</dl>
1271
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1272</div>
1273
1274<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1275<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]1276 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1277</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1278
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1279<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1280
1281<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1282 the GCC "file scope inline asm" blocks. These blocks are internally
1283 concatenated by LLVM and treated as a single unit, but may be separated in
1284 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1285
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1286<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1287module asm "inline asm code goes here"
1288module asm "more can go here"
1289</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1290
1291<p>The strings can contain any character by escaping non-printable characters.
1292 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]1293 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1294
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1295<p>The inline asm code is simply printed to the machine code .s file when
1296 assembly code is generated.</p>
1297
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1298</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1299
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1300<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1301<h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1302 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1303</h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1304
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1305<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1306
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1307<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1308 data is to be laid out in memory. The syntax for the data layout is
1309 simply:</p>
1310
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1311<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1312target datalayout = "<i>layout specification</i>"
1313</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1314
1315<p>The <i>layout specification</i> consists of a list of specifications
1316 separated by the minus sign character ('-'). Each specification starts with
1317 a letter and may include other information after the letter to define some
1318 aspect of the data layout. The specifications accepted are as follows:</p>
1319
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1320<dl>
1321 <dt><tt>E</tt></dt>
1322 <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]1323 bits with the most significance have the lowest address location.</dd>
1324
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1325 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1326 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1327 the bits with the least significance have the lowest address
1328 location.</dd>
1329
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1330 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1331 <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]1332 <i>preferred</i> alignments. All sizes are in bits. Specifying
1333 the <i>pref</i> alignment is optional. If omitted, the
1334 preceding <tt>:</tt> should be omitted too.</dd>
1335
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1336 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1337 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1338 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1339
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1340 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1341 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1342 <i>size</i>.</dd>
1343
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1344 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1345 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +0000[diff] [blame]1346 <i>size</i>. Only values of <i>size</i> that are supported by the target
1347 will work. 32 (float) and 64 (double) are supported on all targets;
1348 80 or 128 (different flavors of long double) are also supported on some
1349 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1350
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1351 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1352 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1353 <i>size</i>.</dd>
1354
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1355 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1356 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1357 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1358
1359 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1360 <dd>This specifies a set of native integer widths for the target CPU
1361 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1362 "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]1363 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1364 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1365</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1366
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1367<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]1368 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1369 specifications in the <tt>datalayout</tt> keyword. The default specifications
1370 are given in this list:</p>
1371
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1372<ul>
1373 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +0000[diff] [blame]1374 <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]1375 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1376 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1377 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1378 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1379 <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]1380 alignment of 64-bits</li>
1381 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1382 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1383 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1384 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1385 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1386 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1387</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1388
1389<p>When LLVM is determining the alignment for a given type, it uses the
1390 following rules:</p>
1391
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1392<ol>
1393 <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]1394 specification is used.</li>
1395
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1396 <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]1397 smallest integer type that is larger than the bitwidth of the sought type
1398 is used. If none of the specifications are larger than the bitwidth then
1399 the the largest integer type is used. For example, given the default
1400 specifications above, the i7 type will use the alignment of i8 (next
1401 largest) while both i65 and i256 will use the alignment of i64 (largest
1402 specified).</li>
1403
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1404 <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]1405 largest vector type that is smaller than the sought vector type will be
1406 used as a fall back. This happens because &lt;128 x double&gt; can be
1407 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1408</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1409
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1410</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1411
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1412<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1413<h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1414 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1415</h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1416
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1417<div>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1418
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1419<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +0000[diff] [blame]1420with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1421is undefined. Pointer values are associated with address ranges
1422according to the following rules:</p>
1423
1424<ul>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1425 <li>A pointer value is associated with the addresses associated with
1426 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1427 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1428 range of the variable's storage.</li>
1429 <li>The result value of an allocation instruction is associated with
1430 the address range of the allocated storage.</li>
1431 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1432 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1433 <li>An integer constant other than zero or a pointer value returned
1434 from a function not defined within LLVM may be associated with address
1435 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1436 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1437 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1438</ul>
1439
1440<p>A pointer value is <i>based</i> on another pointer value according
1441 to the following rules:</p>
1442
1443<ul>
1444 <li>A pointer value formed from a
1445 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1446 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1447 <li>The result value of a
1448 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1449 of the <tt>bitcast</tt>.</li>
1450 <li>A pointer value formed by an
1451 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1452 pointer values that contribute (directly or indirectly) to the
1453 computation of the pointer's value.</li>
1454 <li>The "<i>based</i> on" relationship is transitive.</li>
1455</ul>
1456
1457<p>Note that this definition of <i>"based"</i> is intentionally
1458 similar to the definition of <i>"based"</i> in C99, though it is
1459 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1460
1461<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1462<tt><a href="#i_load">load</a></tt> merely indicates the size and
1463alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +0000[diff] [blame]1464interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1465<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1466and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1467
1468<p>Consequently, type-based alias analysis, aka TBAA, aka
1469<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1470LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1471additional information which specialized optimization passes may use
1472to implement type-based alias analysis.</p>
1473
1474</div>
1475
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1476<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1477<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1478 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1479</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1480
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1481<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1482
1483<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1484href="#i_store"><tt>store</tt></a>s, and <a
1485href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1486The optimizers must not change the number of volatile operations or change their
1487order of execution relative to other volatile operations. The optimizers
1488<i>may</i> change the order of volatile operations relative to non-volatile
1489operations. This is not Java's "volatile" and has no cross-thread
1490synchronization behavior.</p>
1491
1492</div>
1493
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1494<!-- ======================================================================= -->
1495<h3>
1496 <a name="memmodel">Memory Model for Concurrent Operations</a>
1497</h3>
1498
1499<div>
1500
1501<p>The LLVM IR does not define any way to start parallel threads of execution
1502or to register signal handlers. Nonetheless, there are platform-specific
1503ways to create them, and we define LLVM IR's behavior in their presence. This
1504model is inspired by the C++0x memory model.</p>
1505
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1506<p>For a more informal introduction to this model, see the
1507<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.
1508
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1509<p>We define a <i>happens-before</i> partial order as the least partial order
1510that</p>
1511<ul>
1512 <li>Is a superset of single-thread program order, and</li>
1513 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1514 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1515 by platform-specific techniques, like pthread locks, thread
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1516 creation, thread joining, etc., and by atomic instructions.
1517 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1518 </li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1519</ul>
1520
1521<p>Note that program order does not introduce <i>happens-before</i> edges
1522between a thread and signals executing inside that thread.</p>
1523
1524<p>Every (defined) read operation (load instructions, memcpy, atomic
1525loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1526(defined) write operations (store instructions, atomic
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1527stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1528initialized globals are considered to have a write of the initializer which is
1529atomic and happens before any other read or write of the memory in question.
1530For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1531any write to the same byte, except:</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1532
1533<ul>
1534 <li>If <var>write<sub>1</sub></var> happens before
1535 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1536 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1537 does not see <var>write<sub>1</sub></var>.
Bill Wendling0246bb72011-07-31 06:45:03 +0000[diff] [blame]1538 <li>If <var>R<sub>byte</sub></var> happens before
1539 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1540 see <var>write<sub>3</sub></var>.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1541</ul>
1542
1543<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1544<ul>
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1545 <li>If <var>R</var> is volatile, the result is target-dependent. (Volatile
1546 is supposed to give guarantees which can support
1547 <code>sig_atomic_t</code> in C/C++, and may be used for accesses to
1548 addresses which do not behave like normal memory. It does not generally
1549 provide cross-thread synchronization.)
1550 <li>Otherwise, if there is no write to the same byte that happens before
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1551 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1552 <tt>undef</tt> for that byte.
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1553 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1554 <var>R<sub>byte</sub></var> returns the value written by that
1555 write.</li>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1556 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1557 <var>R<sub>byte</sub></var> may see are atomic, it chooses one of the
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1558 values written. See the <a href="#ordering">Atomic Memory Ordering
1559 Constraints</a> section for additional constraints on how the choice
1560 is made.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1561 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1562</ul>
1563
1564<p><var>R</var> returns the value composed of the series of bytes it read.
1565This implies that some bytes within the value may be <tt>undef</tt>
1566<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1567defines the semantics of the operation; it doesn't mean that targets will
1568emit more than one instruction to read the series of bytes.</p>
1569
1570<p>Note that in cases where none of the atomic intrinsics are used, this model
1571places only one restriction on IR transformations on top of what is required
1572for single-threaded execution: introducing a store to a byte which might not
Eli Friedman101c81d2011-08-02 01:15:34 +0000[diff] [blame]1573otherwise be stored is not allowed in general. (Specifically, in the case
1574where another thread might write to and read from an address, introducing a
1575store can change a load that may see exactly one write into a load that may
1576see multiple writes.)</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1577
1578<!-- FIXME: This model assumes all targets where concurrency is relevant have
1579a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1580none of the backends currently in the tree fall into this category; however,
1581there might be targets which care. If there are, we want a paragraph
1582like the following:
1583
1584Targets may specify that stores narrower than a certain width are not
1585available; on such a target, for the purposes of this model, treat any
1586non-atomic write with an alignment or width less than the minimum width
1587as if it writes to the relevant surrounding bytes.
1588-->
1589
1590</div>
1591
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1592<!-- ======================================================================= -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1593<h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1594 <a name="ordering">Atomic Memory Ordering Constraints</a>
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1595</h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1596
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1597<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1598
1599<p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]1600<a href="#i_atomicrmw"><code>atomicrmw</code></a>,
1601<a href="#i_fence"><code>fence</code></a>,
1602<a href="#i_load"><code>atomic load</code></a>, and
Eli Friedman8fa281a2011-08-09 23:26:12 +0000[diff] [blame]1603<a href="#i_store"><code>atomic store</code></a>) take an ordering parameter
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1604that determines which other atomic instructions on the same address they
1605<i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1606but are somewhat more colloquial. If these descriptions aren't precise enough,
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1607check those specs (see spec references in the
1608<a href="Atomic.html#introduction">atomics guide</a>).
1609<a href="#i_fence"><code>fence</code></a> instructions
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1610treat these orderings somewhat differently since they don't take an address.
1611See that instruction's documentation for details.</p>
1612
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1613<p>For a simpler introduction to the ordering constraints, see the
1614<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.</p>
1615
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1616<dl>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1617<dt><code>unordered</code></dt>
1618<dd>The set of values that can be read is governed by the happens-before
1619partial order. A value cannot be read unless some operation wrote it.
1620This is intended to provide a guarantee strong enough to model Java's
1621non-volatile shared variables. This ordering cannot be specified for
1622read-modify-write operations; it is not strong enough to make them atomic
1623in any interesting way.</dd>
1624<dt><code>monotonic</code></dt>
1625<dd>In addition to the guarantees of <code>unordered</code>, there is a single
1626total order for modifications by <code>monotonic</code> operations on each
1627address. All modification orders must be compatible with the happens-before
1628order. There is no guarantee that the modification orders can be combined to
1629a global total order for the whole program (and this often will not be
1630possible). The read in an atomic read-modify-write operation
1631(<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1632<a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1633reads the value in the modification order immediately before the value it
1634writes. If one atomic read happens before another atomic read of the same
1635address, the later read must see the same value or a later value in the
1636address's modification order. This disallows reordering of
1637<code>monotonic</code> (or stronger) operations on the same address. If an
1638address is written <code>monotonic</code>ally by one thread, and other threads
1639<code>monotonic</code>ally read that address repeatedly, the other threads must
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1640eventually see the write. This corresponds to the C++0x/C1x
1641<code>memory_order_relaxed</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1642<dt><code>acquire</code></dt>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1643<dd>In addition to the guarantees of <code>monotonic</code>,
Eli Friedmanc264b2f2011-08-24 20:28:39 +0000[diff] [blame]1644a <i>synchronizes-with</i> edge may be formed with a <code>release</code>
1645operation. This is intended to model C++'s <code>memory_order_acquire</code>.</dd>
1646<dt><code>release</code></dt>
1647<dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1648writes a value which is subsequently read by an <code>acquire</code> operation,
1649it <i>synchronizes-with</i> that operation. (This isn't a complete
1650description; see the C++0x definition of a release sequence.) This corresponds
1651to the C++0x/C1x <code>memory_order_release</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1652<dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1653<code>acquire</code> and <code>release</code> operation on its address.
1654This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1655<dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1656<dd>In addition to the guarantees of <code>acq_rel</code>
1657(<code>acquire</code> for an operation which only reads, <code>release</code>
1658for an operation which only writes), there is a global total order on all
1659sequentially-consistent operations on all addresses, which is consistent with
1660the <i>happens-before</i> partial order and with the modification orders of
1661all the affected addresses. Each sequentially-consistent read sees the last
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1662preceding write to the same address in this global order. This corresponds
1663to the C++0x/C1x <code>memory_order_seq_cst</code> and Java volatile.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1664</dl>
1665
1666<p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1667it only <i>synchronizes with</i> or participates in modification and seq_cst
1668total orderings with other operations running in the same thread (for example,
1669in signal handlers).</p>
1670
1671</div>
1672
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1673</div>
1674
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1675<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1676<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1677<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1678
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1679<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1680
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1681<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1682 intermediate representation. Being typed enables a number of optimizations
1683 to be performed on the intermediate representation directly, without having
1684 to do extra analyses on the side before the transformation. A strong type
1685 system makes it easier to read the generated code and enables novel analyses
1686 and transformations that are not feasible to perform on normal three address
1687 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1688
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1689<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1690<h3>
1691 <a name="t_classifications">Type Classifications</a>
1692</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1693
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1694<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1695
1696<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1697
1698<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1699 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1700 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1701 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1702 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1703 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1704 </tr>
1705 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1706 <td><a href="#t_floating">floating point</a></td>
1707 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1708 </tr>
1709 <tr>
1710 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1711 <td><a href="#t_integer">integer</a>,
1712 <a href="#t_floating">floating point</a>,
1713 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +0000[diff] [blame]1714 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]1715 <a href="#t_struct">structure</a>,
1716 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1717 <a href="#t_label">label</a>,
1718 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1719 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1720 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1721 <tr>
1722 <td><a href="#t_primitive">primitive</a></td>
1723 <td><a href="#t_label">label</a>,
1724 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +0000[diff] [blame]1725 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1726 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1727 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1728 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1729 </tr>
1730 <tr>
1731 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1732 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1733 <a href="#t_function">function</a>,
1734 <a href="#t_pointer">pointer</a>,
1735 <a href="#t_struct">structure</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1736 <a href="#t_vector">vector</a>,
1737 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +0000[diff] [blame]1738 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1739 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1740 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1741</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1742
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1743<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1744 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1745 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1746
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1747</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1748
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1749<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1750<h3>
1751 <a name="t_primitive">Primitive Types</a>
1752</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1753
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1754<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1755
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1756<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1757 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1758
1759<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1760<h4>
1761 <a name="t_integer">Integer Type</a>
1762</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1763
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1764<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1765
1766<h5>Overview:</h5>
1767<p>The integer type is a very simple type that simply specifies an arbitrary
1768 bit width for the integer type desired. Any bit width from 1 bit to
1769 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1770
1771<h5>Syntax:</h5>
1772<pre>
1773 iN
1774</pre>
1775
1776<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1777 value.</p>
1778
1779<h5>Examples:</h5>
1780<table class="layout">
1781 <tr class="layout">
1782 <td class="left"><tt>i1</tt></td>
1783 <td class="left">a single-bit integer.</td>
1784 </tr>
1785 <tr class="layout">
1786 <td class="left"><tt>i32</tt></td>
1787 <td class="left">a 32-bit integer.</td>
1788 </tr>
1789 <tr class="layout">
1790 <td class="left"><tt>i1942652</tt></td>
1791 <td class="left">a really big integer of over 1 million bits.</td>
1792 </tr>
1793</table>
1794
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1795</div>
1796
1797<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1798<h4>
1799 <a name="t_floating">Floating Point Types</a>
1800</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1801
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1802<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1803
1804<table>
1805 <tbody>
1806 <tr><th>Type</th><th>Description</th></tr>
1807 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1808 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1809 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1810 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1811 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1812 </tbody>
1813</table>
1814
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1815</div>
1816
1817<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1818<h4>
1819 <a name="t_x86mmx">X86mmx Type</a>
1820</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1821
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1822<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1823
1824<h5>Overview:</h5>
1825<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>
1826
1827<h5>Syntax:</h5>
1828<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +0000[diff] [blame]1829 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1830</pre>
1831
1832</div>
1833
1834<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1835<h4>
1836 <a name="t_void">Void Type</a>
1837</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1838
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1839<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1840
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1841<h5>Overview:</h5>
1842<p>The void type does not represent any value and has no size.</p>
1843
1844<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1845<pre>
1846 void
1847</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1848
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1849</div>
1850
1851<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1852<h4>
1853 <a name="t_label">Label Type</a>
1854</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1855
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1856<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1857
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1858<h5>Overview:</h5>
1859<p>The label type represents code labels.</p>
1860
1861<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1862<pre>
1863 label
1864</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1865
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1866</div>
1867
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1868<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1869<h4>
1870 <a name="t_metadata">Metadata Type</a>
1871</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1872
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1873<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1874
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1875<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1876<p>The metadata type represents embedded metadata. No derived types may be
1877 created from metadata except for <a href="#t_function">function</a>
1878 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1879
1880<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1881<pre>
1882 metadata
1883</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1884
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1885</div>
1886
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1887</div>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1888
1889<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1890<h3>
1891 <a name="t_derived">Derived Types</a>
1892</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1893
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1894<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1895
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1896<p>The real power in LLVM comes from the derived types in the system. This is
1897 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1898 useful types. Each of these types contain one or more element types which
1899 may be a primitive type, or another derived type. For example, it is
1900 possible to have a two dimensional array, using an array as the element type
1901 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1902
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1903</div>
1904
1905
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1906<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1907<h4>
1908 <a name="t_aggregate">Aggregate Types</a>
1909</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1910
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1911<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1912
1913<p>Aggregate Types are a subset of derived types that can contain multiple
1914 member types. <a href="#t_array">Arrays</a>,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]1915 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1916 aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1917
1918</div>
1919
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1920<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1921<h4>
1922 <a name="t_array">Array Type</a>
1923</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1924
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1925<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1926
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1927<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1928<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1929 sequentially in memory. The array type requires a size (number of elements)
1930 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1931
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1932<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1933<pre>
1934 [&lt;# elements&gt; x &lt;elementtype&gt;]
1935</pre>
1936
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1937<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1938 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1939
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1940<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1941<table class="layout">
1942 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1943 <td class="left"><tt>[40 x i32]</tt></td>
1944 <td class="left">Array of 40 32-bit integer values.</td>
1945 </tr>
1946 <tr class="layout">
1947 <td class="left"><tt>[41 x i32]</tt></td>
1948 <td class="left">Array of 41 32-bit integer values.</td>
1949 </tr>
1950 <tr class="layout">
1951 <td class="left"><tt>[4 x i8]</tt></td>
1952 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1953 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1954</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1955<p>Here are some examples of multidimensional arrays:</p>
1956<table class="layout">
1957 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1958 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1959 <td class="left">3x4 array of 32-bit integer values.</td>
1960 </tr>
1961 <tr class="layout">
1962 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1963 <td class="left">12x10 array of single precision floating point values.</td>
1964 </tr>
1965 <tr class="layout">
1966 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1967 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1968 </tr>
1969</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1970
Dan Gohman7657f6b2009-11-09 19:01:53 +0000[diff] [blame]1971<p>There is no restriction on indexing beyond the end of the array implied by
1972 a static type (though there are restrictions on indexing beyond the bounds
1973 of an allocated object in some cases). This means that single-dimension
1974 'variable sized array' addressing can be implemented in LLVM with a zero
1975 length array type. An implementation of 'pascal style arrays' in LLVM could
1976 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1977
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1978</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1979
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1980<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1981<h4>
1982 <a name="t_function">Function Type</a>
1983</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1984
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1985<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1986
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1987<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1988<p>The function type can be thought of as a function signature. It consists of
1989 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]1990 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]1991
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1992<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1993<pre>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]1994 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1995</pre>
1996
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1997<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]1998 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1999 which indicates that the function takes a variable number of arguments.
2000 Variable argument functions can access their arguments with
2001 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2002 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]2003 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2004
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2005<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2006<table class="layout">
2007 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2008 <td class="left"><tt>i32 (i32)</tt></td>
2009 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2010 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2011 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2012 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]2013 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2014 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2015 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
2016 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2017 </td>
2018 </tr><tr class="layout">
2019 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2020 <td class="left">A vararg function that takes at least one
2021 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2022 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2023 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2024 </td>
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2025 </tr><tr class="layout">
2026 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2027 <td class="left">A function taking an <tt>i32</tt>, returning a
2028 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2029 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2030 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2031</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2032
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2033</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2034
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2035<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2036<h4>
2037 <a name="t_struct">Structure Type</a>
2038</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2039
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2040<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2041
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2042<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2043<p>The structure type is used to represent a collection of data members together
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2044 in memory. The elements of a structure may be any type that has a size.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2045
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]2046<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2047 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2048 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2049 Structures in registers are accessed using the
2050 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2051 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2052
2053<p>Structures may optionally be "packed" structures, which indicate that the
2054 alignment of the struct is one byte, and that there is no padding between
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2055 the elements. In non-packed structs, padding between field types is inserted
2056 as defined by the TargetData string in the module, which is required to match
2057 what the underlying processor expects.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2058
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2059<p>Structures can either be "literal" or "identified". A literal structure is
2060 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) whereas identified
2061 types are always defined at the top level with a name. Literal types are
2062 uniqued by their contents and can never be recursive or opaque since there is
Chris Lattneraa175c32011-08-12 18:12:40 +0000[diff] [blame]2063 no way to write one. Identified types can be recursive, can be opaqued, and are
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2064 never uniqued.
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2065</p>
2066
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2067<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2068<pre>
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2069 %T1 = type { &lt;type list&gt; } <i>; Identified normal struct type</i>
2070 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Identified packed struct type</i>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2071</pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2072
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2073<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2074<table class="layout">
2075 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2076 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2077 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2078 </tr>
2079 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2080 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
2081 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2082 second element is a <a href="#t_pointer">pointer</a> to a
2083 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2084 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2085 </tr>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2086 <tr class="layout">
2087 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
2088 <td class="left">A packed struct known to be 5 bytes in size.</td>
2089 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2090</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2091
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2092</div>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2093
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2094<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2095<h4>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2096 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2097</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2098
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2099<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2100
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2101<h5>Overview:</h5>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2102<p>Opaque structure types are used to represent named structure types that do
2103 not have a body specified. This corresponds (for example) to the C notion of
2104 a forward declared structure.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2105
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2106<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2107<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2108 %X = type opaque
2109 %52 = type opaque
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2110</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2111
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2112<h5>Examples:</h5>
2113<table class="layout">
2114 <tr class="layout">
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2115 <td class="left"><tt>opaque</tt></td>
2116 <td class="left">An opaque type.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2117 </tr>
2118</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2119
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2120</div>
2121
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2122
2123
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2124<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2125<h4>
2126 <a name="t_pointer">Pointer Type</a>
2127</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2128
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2129<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2130
2131<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +0000[diff] [blame]2132<p>The pointer type is used to specify memory locations.
2133 Pointers are commonly used to reference objects in memory.</p>
2134
2135<p>Pointer types may have an optional address space attribute defining the
2136 numbered address space where the pointed-to object resides. The default
2137 address space is number zero. The semantics of non-zero address
2138 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2139
2140<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2141 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2142
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2143<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2144<pre>
2145 &lt;type&gt; *
2146</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2147
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2148<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2149<table class="layout">
2150 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2151 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2152 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2153 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2154 </tr>
2155 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]2156 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2157 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2158 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2159 <tt>i32</tt>.</td>
2160 </tr>
2161 <tr class="layout">
2162 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2163 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2164 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2165 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2166</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2167
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2168</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2169
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2170<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2171<h4>
2172 <a name="t_vector">Vector Type</a>
2173</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2174
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2175<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2176
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2177<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2178<p>A vector type is a simple derived type that represents a vector of elements.
2179 Vector types are used when multiple primitive data are operated in parallel
2180 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +0000[diff] [blame]2181 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2182 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2183
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2184<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2185<pre>
2186 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2187</pre>
2188
Chris Lattner7d2e7be2010-10-10 18:20:35 +0000[diff] [blame]2189<p>The number of elements is a constant integer value larger than 0; elementtype
2190 may be any integer or floating point type. Vectors of size zero are not
2191 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2192
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2193<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2194<table class="layout">
2195 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2196 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2197 <td class="left">Vector of 4 32-bit integer values.</td>
2198 </tr>
2199 <tr class="layout">
2200 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2201 <td class="left">Vector of 8 32-bit floating-point values.</td>
2202 </tr>
2203 <tr class="layout">
2204 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2205 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2206 </tr>
2207</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2208
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2209</div>
2210
Bill Wendlingaf75f0c2011-07-31 06:47:33 +0000[diff] [blame]2211</div>
2212
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2213<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2214<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2215<!-- *********************************************************************** -->
2216
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2217<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2218
2219<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2220 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2221
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2222<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2223<h3>
2224 <a name="simpleconstants">Simple Constants</a>
2225</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2226
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2227<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2228
2229<dl>
2230 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2231 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]2232 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2233
2234 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2235 <dd>Standard integers (such as '4') are constants of
2236 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2237 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2238
2239 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2240 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2241 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2242 notation (see below). The assembler requires the exact decimal value of a
2243 floating-point constant. For example, the assembler accepts 1.25 but
2244 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2245 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2246
2247 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2248 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2249 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2250</dl>
2251
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2252<p>The one non-intuitive notation for constants is the hexadecimal form of
2253 floating point constants. For example, the form '<tt>double
2254 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2255 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2256 constants are required (and the only time that they are generated by the
2257 disassembler) is when a floating point constant must be emitted but it cannot
2258 be represented as a decimal floating point number in a reasonable number of
2259 digits. For example, NaN's, infinities, and other special values are
2260 represented in their IEEE hexadecimal format so that assembly and disassembly
2261 do not cause any bits to change in the constants.</p>
2262
Dale Johannesenbd5e5a82009-02-11 22:14:51 +0000[diff] [blame]2263<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2264 represented using the 16-digit form shown above (which matches the IEEE754
2265 representation for double); float values must, however, be exactly
2266 representable as IEE754 single precision. Hexadecimal format is always used
2267 for long double, and there are three forms of long double. The 80-bit format
2268 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2269 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2270 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2271 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2272 currently supported target uses this format. Long doubles will only work if
2273 they match the long double format on your target. All hexadecimal formats
2274 are big-endian (sign bit at the left).</p>
2275
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]2276<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2277</div>
2278
2279<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2280<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]2281<a name="aggregateconstants"></a> <!-- old anchor -->
2282<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2283</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2284
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2285<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2286
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2287<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2288 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2289
2290<dl>
2291 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2292 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2293 type definitions (a comma separated list of elements, surrounded by braces
2294 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2295 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2296 Structure constants must have <a href="#t_struct">structure type</a>, and
2297 the number and types of elements must match those specified by the
2298 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2299
2300 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2301 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2302 definitions (a comma separated list of elements, surrounded by square
2303 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2304 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2305 the number and types of elements must match those specified by the
2306 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2307
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2308 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2309 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2310 definitions (a comma separated list of elements, surrounded by
2311 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2312 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2313 have <a href="#t_vector">vector type</a>, and the number and types of
2314 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2315
2316 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2317 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]2318 value to zero of <em>any</em> type, including scalar and
2319 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2320 This is often used to avoid having to print large zero initializers
2321 (e.g. for large arrays) and is always exactly equivalent to using explicit
2322 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +0000[diff] [blame]2323
2324 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +0000[diff] [blame]2325 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2326 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2327 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2328 be interpreted as part of the instruction stream, metadata is a place to
2329 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2330</dl>
2331
2332</div>
2333
2334<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2335<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2336 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2337</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2338
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2339<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2340
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2341<p>The addresses of <a href="#globalvars">global variables</a>
2342 and <a href="#functionstructure">functions</a> are always implicitly valid
2343 (link-time) constants. These constants are explicitly referenced when
2344 the <a href="#identifiers">identifier for the global</a> is used and always
2345 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2346 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2347
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2348<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]2349@X = global i32 17
2350@Y = global i32 42
2351@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2352</pre>
2353
2354</div>
2355
2356<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2357<h3>
2358 <a name="undefvalues">Undefined Values</a>
2359</h3>
2360
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2361<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2362
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2363<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]2364 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2365 Undefined values may be of any type (other than '<tt>label</tt>'
2366 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2367
Chris Lattnerc608cb12009-09-11 01:49:31 +0000[diff] [blame]2368<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2369 program is well defined no matter what value is used. This gives the
2370 compiler more freedom to optimize. Here are some examples of (potentially
2371 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2372
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2373
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2374<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2375 %A = add %X, undef
2376 %B = sub %X, undef
2377 %C = xor %X, undef
2378Safe:
2379 %A = undef
2380 %B = undef
2381 %C = undef
2382</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2383
2384<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]2385 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]2386
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2387<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2388 %A = or %X, undef
2389 %B = and %X, undef
2390Safe:
2391 %A = -1
2392 %B = 0
2393Unsafe:
2394 %A = undef
2395 %B = undef
2396</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2397
2398<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2399 For example, if <tt>%X</tt> has a zero bit, then the output of the
2400 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2401 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2402 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2403 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2404 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2405 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2406 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2407
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2408<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2409 %A = select undef, %X, %Y
2410 %B = select undef, 42, %Y
2411 %C = select %X, %Y, undef
2412Safe:
2413 %A = %X (or %Y)
2414 %B = 42 (or %Y)
2415 %C = %Y
2416Unsafe:
2417 %A = undef
2418 %B = undef
2419 %C = undef
2420</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2421
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2422<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2423 branch) conditions can go <em>either way</em>, but they have to come from one
2424 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2425 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2426 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2427 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2428 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2429 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2430
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2431<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2432 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2433
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2434 %B = undef
2435 %C = xor %B, %B
2436
2437 %D = undef
2438 %E = icmp lt %D, 4
2439 %F = icmp gte %D, 4
2440
2441Safe:
2442 %A = undef
2443 %B = undef
2444 %C = undef
2445 %D = undef
2446 %E = undef
2447 %F = undef
2448</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2449
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2450<p>This example points out that two '<tt>undef</tt>' operands are not
2451 necessarily the same. This can be surprising to people (and also matches C
2452 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2453 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2454 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2455 its value over its "live range". This is true because the variable doesn't
2456 actually <em>have a live range</em>. Instead, the value is logically read
2457 from arbitrary registers that happen to be around when needed, so the value
2458 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2459 need to have the same semantics or the core LLVM "replace all uses with"
2460 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2461
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2462<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2463 %A = fdiv undef, %X
2464 %B = fdiv %X, undef
2465Safe:
2466 %A = undef
2467b: unreachable
2468</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2469
2470<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2471 value</em> and <em>undefined behavior</em>. An undefined value (like
2472 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2473 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2474 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2475 defined on SNaN's. However, in the second example, we can make a more
2476 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2477 arbitrary value, we are allowed to assume that it could be zero. Since a
2478 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2479 the operation does not execute at all. This allows us to delete the divide and
2480 all code after it. Because the undefined operation "can't happen", the
2481 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2482
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2483<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2484a: store undef -> %X
2485b: store %X -> undef
2486Safe:
2487a: &lt;deleted&gt;
2488b: unreachable
2489</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2490
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2491<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2492 undefined value can be assumed to not have any effect; we can assume that the
2493 value is overwritten with bits that happen to match what was already there.
2494 However, a store <em>to</em> an undefined location could clobber arbitrary
2495 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2496
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2497</div>
2498
2499<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2500<h3>
2501 <a name="trapvalues">Trap Values</a>
2502</h3>
2503
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2504<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2505
Dan Gohmanc68ce062010-04-26 20:21:21 +0000[diff] [blame]2506<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2507 instead of representing an unspecified bit pattern, they represent the
2508 fact that an instruction or constant expression which cannot evoke side
2509 effects has nevertheless detected a condition which results in undefined
Dan Gohmanc68ce062010-04-26 20:21:21 +0000[diff] [blame]2510 behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2511
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2512<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]2513 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2514 <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]2515
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2516<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2517
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2518<ul>
2519<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2520 their operands.</li>
2521
2522<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2523 to their dynamic predecessor basic block.</li>
2524
2525<li>Function arguments depend on the corresponding actual argument values in
2526 the dynamic callers of their functions.</li>
2527
2528<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2529 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2530 control back to them.</li>
2531
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2532<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2533 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2534 or exception-throwing call instructions that dynamically transfer control
2535 back to them.</li>
2536
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2537<li>Non-volatile loads and stores depend on the most recent stores to all of the
2538 referenced memory addresses, following the order in the IR
2539 (including loads and stores implied by intrinsics such as
2540 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2541
Dan Gohman7c24ff12010-05-03 14:59:34 +0000[diff] [blame]2542<!-- TODO: In the case of multiple threads, this only applies if the store
2543 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2544
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2545<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2546
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2547<li>An instruction with externally visible side effects depends on the most
2548 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]2549 the order in the IR. (This includes
2550 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2551
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2552<li>An instruction <i>control-depends</i> on a
2553 <a href="#terminators">terminator instruction</a>
2554 if the terminator instruction has multiple successors and the instruction
2555 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2556 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2557
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2558<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2559 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2560 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2561 successor.</li>
2562
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2563<li>Dependence is transitive.</li>
2564
2565</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2566
2567<p>Whenever a trap value is generated, all values which depend on it evaluate
2568 to trap. If they have side effects, the evoke their side effects as if each
2569 operand with a trap value were undef. If they have externally-visible side
2570 effects, the behavior is undefined.</p>
2571
2572<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +0000[diff] [blame]2573
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2574<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2575entry:
2576 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2577 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2578 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2579 store i32 0, i32* %trap_yet_again ; undefined behavior
2580
2581 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2582 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2583
2584 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2585
2586 %narrowaddr = bitcast i32* @g to i16*
2587 %wideaddr = bitcast i32* @g to i64*
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2588 %trap3 = load i16* %narrowaddr ; Returns a trap value.
2589 %trap4 = load i64* %wideaddr ; Returns a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2590
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2591 %cmp = icmp slt i32 %trap, 0 ; Returns a trap value.
2592 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2593
2594true:
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2595 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2596 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2597 br label %end
2598
2599end:
2600 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2601 ; Both edges into this PHI are
2602 ; control-dependent on %cmp, so this
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2603 ; always results in a trap value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2604
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2605 volatile store i32 0, i32* @g ; This would depend on the store in %true
2606 ; if %cmp is true, or the store in %entry
2607 ; otherwise, so this is undefined behavior.
2608
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2609 br i1 %cmp, label %second_true, label %second_end
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2610 ; The same branch again, but this time the
2611 ; true block doesn't have side effects.
2612
2613second_true:
2614 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2615 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2616
2617second_end:
2618 volatile store i32 0, i32* @g ; This time, the instruction always depends
2619 ; on the store in %end. Also, it is
2620 ; control-equivalent to %end, so this is
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2621 ; well-defined (again, ignoring earlier
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2622 ; undefined behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2623</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2624
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2625</div>
2626
2627<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2628<h3>
2629 <a name="blockaddress">Addresses of Basic Blocks</a>
2630</h3>
2631
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2632<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2633
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2634<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2635
2636<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +0000[diff] [blame]2637 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2638 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2639
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2640<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2641 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2642 comparisons against null. Pointer equality tests between labels addresses
2643 results in undefined behavior &mdash; though, again, comparison against null
2644 is ok, and no label is equal to the null pointer. This may be passed around
2645 as an opaque pointer sized value as long as the bits are not inspected. This
2646 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2647 long as the original value is reconstituted before the <tt>indirectbr</tt>
2648 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2649
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2650<p>Finally, some targets may provide defined semantics when using the value as
2651 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2652
2653</div>
2654
2655
2656<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2657<h3>
2658 <a name="constantexprs">Constant Expressions</a>
2659</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2660
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2661<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2662
2663<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2664 to be used as constants. Constant expressions may be of
2665 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2666 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2667 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2668
2669<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2670 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2671 <dd>Truncate a constant to another type. The bit size of CST must be larger
2672 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2673
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2674 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2675 <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]2676 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2677
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2678 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2679 <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]2680 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2681
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2682 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2683 <dd>Truncate a floating point constant to another floating point type. The
2684 size of CST must be larger than the size of TYPE. Both types must be
2685 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2686
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2687 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2688 <dd>Floating point extend a constant to another type. The size of CST must be
2689 smaller or equal to the size of TYPE. Both types must be floating
2690 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2691
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2692 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2693 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2694 constant. TYPE must be a scalar or vector integer type. CST must be of
2695 scalar or vector floating point type. Both CST and TYPE must be scalars,
2696 or vectors of the same number of elements. If the value won't fit in the
2697 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2698
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2699 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2700 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2701 constant. TYPE must be a scalar or vector integer type. CST must be of
2702 scalar or vector floating point type. Both CST and TYPE must be scalars,
2703 or vectors of the same number of elements. If the value won't fit in the
2704 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2705
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2706 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2707 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2708 constant. TYPE must be a scalar or vector floating point type. CST must be
2709 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2710 vectors of the same number of elements. If the value won't fit in the
2711 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2712
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2713 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2714 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2715 constant. TYPE must be a scalar or vector floating point type. CST must be
2716 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2717 vectors of the same number of elements. If the value won't fit in the
2718 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2719
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2720 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2721 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2722 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2723 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2724 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2725
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2726 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2727 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2728 type. CST must be of integer type. The CST value is zero extended,
2729 truncated, or unchanged to make it fit in a pointer size. This one is
2730 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2731
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2732 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +0000[diff] [blame]2733 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2734 are the same as those for the <a href="#i_bitcast">bitcast
2735 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2736
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2737 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2738 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2739 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2740 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2741 instruction, the index list may have zero or more indexes, which are
2742 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2743
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2744 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2745 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2746
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2747 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2748 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2749
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2750 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2751 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2752
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2753 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2754 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2755 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2756
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2757 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2758 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2759 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2760
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2761 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2762 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2763 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2764
Nick Lewycky9e130ce2010-05-29 06:44:15 +0000[diff] [blame]2765 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2766 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2767 constants. The index list is interpreted in a similar manner as indices in
2768 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2769 index value must be specified.</dd>
2770
2771 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2772 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2773 constants. The index list is interpreted in a similar manner as indices in
2774 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2775 index value must be specified.</dd>
2776
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2777 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2778 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2779 be any of the <a href="#binaryops">binary</a>
2780 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2781 on operands are the same as those for the corresponding instruction
2782 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2783</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2784
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2785</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]2786
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2787</div>
2788
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2789<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2790<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2791<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2792<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2793<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2794<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2795<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2796</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2797
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2798<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2799
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2800<p>LLVM supports inline assembler expressions (as opposed
2801 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2802 a special value. This value represents the inline assembler as a string
2803 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2804 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2805 expression has side effects, and a flag indicating whether the function
2806 containing the asm needs to align its stack conservatively. An example
2807 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2808
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2809<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2810i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2811</pre>
2812
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2813<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2814 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2815 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2816
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2817<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2818%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]2819</pre>
2820
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2821<p>Inline asms with side effects not visible in the constraint list must be
2822 marked as having side effects. This is done through the use of the
2823 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2824
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2825<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2826call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2827</pre>
2828
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2829<p>In some cases inline asms will contain code that will not work unless the
2830 stack is aligned in some way, such as calls or SSE instructions on x86,
2831 yet will not contain code that does that alignment within the asm.
2832 The compiler should make conservative assumptions about what the asm might
2833 contain and should generate its usual stack alignment code in the prologue
2834 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2835
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2836<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2837call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2838</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2839
2840<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2841 first.</p>
2842
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2843<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2844 documented here. Constraints on what can be done (e.g. duplication, moving,
2845 etc need to be documented). This is probably best done by reference to
2846 another document that covers inline asm from a holistic perspective.</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2847
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2848<h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2849<a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2850</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2851
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2852<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2853
2854<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2855 attached to it that contains a list of constant integers. If present, the
2856 code generator will use the integer as the location cookie value when report
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2857 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]2858 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2859 source code that produced it. For example:</p>
2860
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2861<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2862call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2863...
2864!42 = !{ i32 1234567 }
2865</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2866
2867<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]2868 IR. If the MDNode contains multiple constants, the code generator will use
2869 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]2870
2871</div>
2872
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2873</div>
2874
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2875<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2876<h3>
2877 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2878</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2879
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2880<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2881
2882<p>LLVM IR allows metadata to be attached to instructions in the program that
2883 can convey extra information about the code to the optimizers and code
2884 generator. One example application of metadata is source-level debug
2885 information. There are two metadata primitives: strings and nodes. All
2886 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2887 preceding exclamation point ('<tt>!</tt>').</p>
2888
2889<p>A metadata string is a string surrounded by double quotes. It can contain
2890 any character by escaping non-printable characters with "\xx" where "xx" is
2891 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2892
2893<p>Metadata nodes are represented with notation similar to structure constants
2894 (a comma separated list of elements, surrounded by braces and preceded by an
2895 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2896 10}</tt>". Metadata nodes can have any values as their operand.</p>
2897
2898<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2899 metadata nodes, which can be looked up in the module symbol table. For
2900 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2901
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2902<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2903 function is using two metadata arguments.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2904
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2905<div class="doc_code">
2906<pre>
2907call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2908</pre>
2909</div>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2910
2911<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2912 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2913
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2914<div class="doc_code">
2915<pre>
2916%indvar.next = add i64 %indvar, 1, !dbg !21
2917</pre>
2918</div>
2919
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2920</div>
2921
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2922</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2923
2924<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2925<h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2926 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2927</h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2928<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2929<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2930<p>LLVM has a number of "magic" global variables that contain data that affect
2931code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2932of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2933section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2934by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2935
2936<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2937<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2938<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2939</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2940
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2941<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2942
2943<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2944href="#linkage_appending">appending linkage</a>. This array contains a list of
2945pointers to global variables and functions which may optionally have a pointer
2946cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2947
2948<pre>
2949 @X = global i8 4
2950 @Y = global i32 123
2951
2952 @llvm.used = appending global [2 x i8*] [
2953 i8* @X,
2954 i8* bitcast (i32* @Y to i8*)
2955 ], section "llvm.metadata"
2956</pre>
2957
2958<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2959compiler, assembler, and linker are required to treat the symbol as if there is
2960a reference to the global that it cannot see. For example, if a variable has
2961internal linkage and no references other than that from the <tt>@llvm.used</tt>
2962list, it cannot be deleted. This is commonly used to represent references from
2963inline asms and other things the compiler cannot "see", and corresponds to
2964"attribute((used))" in GNU C.</p>
2965
2966<p>On some targets, the code generator must emit a directive to the assembler or
2967object file to prevent the assembler and linker from molesting the symbol.</p>
2968
2969</div>
2970
2971<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2972<h3>
2973 <a name="intg_compiler_used">
2974 The '<tt>llvm.compiler.used</tt>' Global Variable
2975 </a>
2976</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2977
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2978<div>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2979
2980<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2981<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2982touching the symbol. On targets that support it, this allows an intelligent
2983linker to optimize references to the symbol without being impeded as it would be
2984by <tt>@llvm.used</tt>.</p>
2985
2986<p>This is a rare construct that should only be used in rare circumstances, and
2987should not be exposed to source languages.</p>
2988
2989</div>
2990
2991<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2992<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2993<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2994</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2995
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2996<div>
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2997<pre>
2998%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]2999@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3000</pre>
3001<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.
3002</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3003
3004</div>
3005
3006<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3007<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3008<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3009</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3010
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3011<div>
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3012<pre>
3013%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3014@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3015</pre>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3016
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3017<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.
3018</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3019
3020</div>
3021
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3022</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3023
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]3024<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3025<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3026<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3027
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3028<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3029
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3030<p>The LLVM instruction set consists of several different classifications of
3031 instructions: <a href="#terminators">terminator
3032 instructions</a>, <a href="#binaryops">binary instructions</a>,
3033 <a href="#bitwiseops">bitwise binary instructions</a>,
3034 <a href="#memoryops">memory instructions</a>, and
3035 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3036
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3037<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3038<h3>
3039 <a name="terminators">Terminator Instructions</a>
3040</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3041
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3042<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3043
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3044<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3045 in a program ends with a "Terminator" instruction, which indicates which
3046 block should be executed after the current block is finished. These
3047 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3048 control flow, not values (the one exception being the
3049 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3050
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3051<p>The terminator instructions are:
3052 '<a href="#i_ret"><tt>ret</tt></a>',
3053 '<a href="#i_br"><tt>br</tt></a>',
3054 '<a href="#i_switch"><tt>switch</tt></a>',
3055 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>',
3056 '<a href="#i_invoke"><tt>invoke</tt></a>',
3057 '<a href="#i_unwind"><tt>unwind</tt></a>',
3058 '<a href="#i_resume"><tt>resume</tt></a>', and
3059 '<a href="#i_unreachable"><tt>unreachable</tt></a>'.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3060
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3061<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3062<h4>
3063 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3064</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3065
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3066<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3067
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3068<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3069<pre>
3070 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]3071 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3072</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3073
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3074<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3075<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3076 a value) from a function back to the caller.</p>
3077
3078<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3079 value and then causes control flow, and one that just causes control flow to
3080 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3081
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3082<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3083<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3084 return value. The type of the return value must be a
3085 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3086
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3087<p>A function is not <a href="#wellformed">well formed</a> if it it has a
3088 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3089 value or a return value with a type that does not match its type, or if it
3090 has a void return type and contains a '<tt>ret</tt>' instruction with a
3091 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3092
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3093<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3094<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3095 the calling function's context. If the caller is a
3096 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3097 instruction after the call. If the caller was an
3098 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3099 the beginning of the "normal" destination block. If the instruction returns
3100 a value, that value shall set the call or invoke instruction's return
3101 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3102
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3103<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3104<pre>
3105 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]3106 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +0000[diff] [blame]3107 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]3108</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]3109
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3110</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3111<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3112<h4>
3113 <a name="i_br">'<tt>br</tt>' Instruction</a>
3114</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3115
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3116<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3117
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3118<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3119<pre>
Bill Wendlingb3aa4712011-07-26 10:41:15 +0000[diff] [blame]3120 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3121 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3122</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3123
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3124<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3125<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3126 different basic block in the current function. There are two forms of this
3127 instruction, corresponding to a conditional branch and an unconditional
3128 branch.</p>
3129
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3130<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3131<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3132 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3133 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3134 target.</p>
3135
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3136<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3137<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3138 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3139 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3140 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3141
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3142<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]3143<pre>
3144Test:
3145 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3146 br i1 %cond, label %IfEqual, label %IfUnequal
3147IfEqual:
3148 <a href="#i_ret">ret</a> i32 1
3149IfUnequal:
3150 <a href="#i_ret">ret</a> i32 0
3151</pre>
3152
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3153</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3154
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3155<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3156<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3157 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3158</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3159
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3160<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3161
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3162<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3163<pre>
3164 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3165</pre>
3166
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3167<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3168<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]3169 several different places. It is a generalization of the '<tt>br</tt>'
3170 instruction, allowing a branch to occur to one of many possible
3171 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3172
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3173<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3174<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3175 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3176 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3177 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3178
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3179<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3180<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3181 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3182 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3183 transferred to the corresponding destination; otherwise, control flow is
3184 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3185
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3186<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3187<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3188 <tt>switch</tt> instruction, this instruction may be code generated in
3189 different ways. For example, it could be generated as a series of chained
3190 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3191
3192<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3193<pre>
3194 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3195 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3196 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3197
3198 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3199 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3200
3201 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3202 switch i32 %val, label %otherwise [ i32 0, label %onzero
3203 i32 1, label %onone
3204 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3205</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3206
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3207</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3208
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3209
3210<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3211<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3212 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3213</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3214
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3215<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3216
3217<h5>Syntax:</h5>
3218<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3219 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]3220</pre>
3221
3222<h5>Overview:</h5>
3223
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3224<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3225 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]3226 "<tt>address</tt>". Address must be derived from a <a
3227 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3228
3229<h5>Arguments:</h5>
3230
3231<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3232 rest of the arguments indicate the full set of possible destinations that the
3233 address may point to. Blocks are allowed to occur multiple times in the
3234 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3235
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3236<p>This destination list is required so that dataflow analysis has an accurate
3237 understanding of the CFG.</p>
3238
3239<h5>Semantics:</h5>
3240
3241<p>Control transfers to the block specified in the address argument. All
3242 possible destination blocks must be listed in the label list, otherwise this
3243 instruction has undefined behavior. This implies that jumps to labels
3244 defined in other functions have undefined behavior as well.</p>
3245
3246<h5>Implementation:</h5>
3247
3248<p>This is typically implemented with a jump through a register.</p>
3249
3250<h5>Example:</h5>
3251<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3252 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3253</pre>
3254
3255</div>
3256
3257
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3258<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3259<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3260 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3261</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3262
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3263<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3264
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3265<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3266<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3267 &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]3268 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3269</pre>
3270
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3271<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3272<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3273 function, with the possibility of control flow transfer to either the
3274 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3275 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3276 control flow will return to the "normal" label. If the callee (or any
3277 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3278 instruction, control is interrupted and continued at the dynamically nearest
3279 "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3280
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3281<p>The '<tt>exception</tt>' label is a
3282 <i><a href="ExceptionHandling.html#overview">landing pad</a></i> for the
3283 exception. As such, '<tt>exception</tt>' label is required to have the
3284 "<a href="#i_landingpad"><tt>landingpad</tt></a>" instruction, which contains
3285 the information about about the behavior of the program after unwinding
3286 happens, as its first non-PHI instruction. The restrictions on the
3287 "<tt>landingpad</tt>" instruction's tightly couples it to the
3288 "<tt>invoke</tt>" instruction, so that the important information contained
3289 within the "<tt>landingpad</tt>" instruction can't be lost through normal
3290 code motion.</p>
3291
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3292<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3293<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3294
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3295<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3296 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3297 convention</a> the call should use. If none is specified, the call
3298 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3299
3300 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3301 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3302 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3303
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3304 <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]3305 function value being invoked. In most cases, this is a direct function
3306 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3307 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3308
3309 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3310 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3311
3312 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]3313 signature argument types and parameter attributes. All arguments must be
3314 of <a href="#t_firstclass">first class</a> type. If the function
3315 signature indicates the function accepts a variable number of arguments,
3316 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3317
3318 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3319 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3320
3321 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3322 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3323
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3324 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3325 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3326 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3327</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3328
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3329<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3330<p>This instruction is designed to operate as a standard
3331 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3332 primary difference is that it establishes an association with a label, which
3333 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3334
3335<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3336 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3337 exception. Additionally, this is important for implementation of
3338 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3339
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3340<p>For the purposes of the SSA form, the definition of the value returned by the
3341 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3342 block to the "normal" label. If the callee unwinds then no return value is
3343 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +0000[diff] [blame]3344
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3345<p>Note that the code generator does not yet completely support unwind, and
3346that the invoke/unwind semantics are likely to change in future versions.</p>
3347
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3348<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3349<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3350 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3351 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3352 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3353 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3354</pre>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3355
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3356</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3357
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3358<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3359
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3360<h4>
3361 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3362</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3363
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3364<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3365
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3366<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3367<pre>
3368 unwind
3369</pre>
3370
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3371<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3372<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3373 at the first callee in the dynamic call stack which used
3374 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3375 This is primarily used to implement exception handling.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3376
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3377<h5>Semantics:</h5>
Chris Lattner72ed2002008-04-19 21:01:16 +0000[diff] [blame]3378<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3379 immediately halt. The dynamic call stack is then searched for the
3380 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3381 Once found, execution continues at the "exceptional" destination block
3382 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3383 instruction in the dynamic call chain, undefined behavior results.</p>
3384
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3385<p>Note that the code generator does not yet completely support unwind, and
3386that the invoke/unwind semantics are likely to change in future versions.</p>
3387
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3388</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3389
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3390 <!-- _______________________________________________________________________ -->
3391
3392<h4>
3393 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3394</h4>
3395
3396<div>
3397
3398<h5>Syntax:</h5>
3399<pre>
3400 resume &lt;type&gt; &lt;value&gt;
3401</pre>
3402
3403<h5>Overview:</h5>
3404<p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3405 successors.</p>
3406
3407<h5>Arguments:</h5>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3408<p>The '<tt>resume</tt>' instruction requires one argument, which must have the
Bill Wendlinge4ad50b2011-08-03 18:37:32 +0000[diff] [blame]3409 same type as the result of any '<tt>landingpad</tt>' instruction in the same
3410 function.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3411
3412<h5>Semantics:</h5>
3413<p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3414 (in-flight) exception whose unwinding was interrupted with
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3415 a <a href="#i_landingpad"><tt>landingpad</tt></a> instruction.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3416
3417<h5>Example:</h5>
3418<pre>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3419 resume { i8*, i32 } %exn
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3420</pre>
3421
3422</div>
3423
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3424<!-- _______________________________________________________________________ -->
3425
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3426<h4>
3427 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3428</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3429
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3430<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3431
3432<h5>Syntax:</h5>
3433<pre>
3434 unreachable
3435</pre>
3436
3437<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3438<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3439 instruction is used to inform the optimizer that a particular portion of the
3440 code is not reachable. This can be used to indicate that the code after a
3441 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3442
3443<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3444<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3445
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3446</div>
3447
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3448</div>
3449
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3450<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3451<h3>
3452 <a name="binaryops">Binary Operations</a>
3453</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3454
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3455<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3456
3457<p>Binary operators are used to do most of the computation in a program. They
3458 require two operands of the same type, execute an operation on them, and
3459 produce a single value. The operands might represent multiple data, as is
3460 the case with the <a href="#t_vector">vector</a> data type. The result value
3461 has the same type as its operands.</p>
3462
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3463<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3464
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3465<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3466<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3467 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3468</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3469
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3470<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3471
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3472<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3473<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3474 &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]3475 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3476 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3477 &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]3478</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3479
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3480<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3481<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3482
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3483<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3484<p>The two arguments to the '<tt>add</tt>' instruction must
3485 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3486 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3487
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3488<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3489<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3490
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3491<p>If the sum has unsigned overflow, the result returned is the mathematical
3492 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]3493
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3494<p>Because LLVM integers use a two's complement representation, this instruction
3495 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3496
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3497<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3498 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3499 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3500 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3501 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3502
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3503<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3504<pre>
3505 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3506</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3507
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3508</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3509
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3510<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3511<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3512 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3513</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3514
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3515<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3516
3517<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3518<pre>
3519 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3520</pre>
3521
3522<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3523<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3524
3525<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3526<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3527 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3528 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3529
3530<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3531<p>The value produced is the floating point sum of the two operands.</p>
3532
3533<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3534<pre>
3535 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3536</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3537
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3538</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3539
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3540<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3541<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3542 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3543</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3544
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3545<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3546
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3547<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3548<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3549 &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]3550 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3551 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3552 &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]3553</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3554
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3555<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3556<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3557 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3558
3559<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3560 '<tt>neg</tt>' instruction present in most other intermediate
3561 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3562
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3563<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3564<p>The two arguments to the '<tt>sub</tt>' instruction must
3565 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3566 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3567
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3568<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3569<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3570
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3571<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3572 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3573 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3574
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3575<p>Because LLVM integers use a two's complement representation, this instruction
3576 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3577
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3578<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3579 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3580 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3581 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3582 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3583
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3584<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]3585<pre>
3586 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3587 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3588</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3589
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3590</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3591
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3592<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3593<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3594 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3595</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3596
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3597<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3598
3599<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3600<pre>
3601 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3602</pre>
3603
3604<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3605<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3606 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3607
3608<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3609 '<tt>fneg</tt>' instruction present in most other intermediate
3610 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3611
3612<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]3613<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3614 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3615 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3616
3617<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3618<p>The value produced is the floating point difference of the two operands.</p>
3619
3620<h5>Example:</h5>
3621<pre>
3622 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3623 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3624</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3625
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3626</div>
3627
3628<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3629<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3630 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3631</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3632
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3633<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3634
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3635<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3636<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3637 &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]3638 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3639 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3640 &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]3641</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3642
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3643<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3644<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3645
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3646<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3647<p>The two arguments to the '<tt>mul</tt>' instruction must
3648 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3649 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3650
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3651<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3652<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3653
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3654<p>If the result of the multiplication has unsigned overflow, the result
3655 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3656 width of the result.</p>
3657
3658<p>Because LLVM integers use a two's complement representation, and the result
3659 is the same width as the operands, this instruction returns the correct
3660 result for both signed and unsigned integers. If a full product
3661 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3662 be sign-extended or zero-extended as appropriate to the width of the full
3663 product.</p>
3664
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3665<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3666 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3667 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3668 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3669 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3670
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3671<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3672<pre>
3673 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3674</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3675
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3676</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3677
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3678<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3679<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3680 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3681</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3682
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3683<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3684
3685<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3686<pre>
3687 &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]3688</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3689
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3690<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3691<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3692
3693<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3694<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3695 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3696 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3697
3698<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3699<p>The value produced is the floating point product of the two operands.</p>
3700
3701<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3702<pre>
3703 &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]3704</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3705
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3706</div>
3707
3708<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3709<h4>
3710 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3711</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3712
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3713<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3714
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3715<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3716<pre>
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3717 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3718 &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]3719</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3720
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3721<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3722<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3723
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3724<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3725<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3726 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3727 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3728
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3729<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3730<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3731
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3732<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3733 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3734
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3735<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3736
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3737<p>If the <tt>exact</tt> keyword is present, the result value of the
3738 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3739 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3740
3741
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3742<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3743<pre>
3744 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3745</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3746
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3747</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3748
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3749<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3750<h4>
3751 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3752</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3753
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3754<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3755
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3756<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3757<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3758 &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]3759 &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]3760</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3761
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3762<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3763<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3764
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3765<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3766<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3767 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3768 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3769
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3770<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3771<p>The value produced is the signed integer quotient of the two operands rounded
3772 towards zero.</p>
3773
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3774<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3775 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3776
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3777<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3778 undefined behavior; this is a rare case, but can occur, for example, by doing
3779 a 32-bit division of -2147483648 by -1.</p>
3780
Dan Gohman9c5beed2009-07-22 00:04:19 +0000[diff] [blame]3781<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]3782 <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]3783 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3784
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3785<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3786<pre>
3787 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3788</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3789
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3790</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3791
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3792<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3793<h4>
3794 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3795</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3796
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3797<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3798
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3799<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3800<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3801 &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]3802</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3803
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3804<h5>Overview:</h5>
3805<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3806
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3807<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3808<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3809 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3810 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3811
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3812<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3813<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3814
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3815<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3816<pre>
3817 &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]3818</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3819
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3820</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3821
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3822<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3823<h4>
3824 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3825</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3826
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3827<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3828
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3829<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3830<pre>
3831 &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]3832</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3833
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3834<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3835<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3836 division of its two arguments.</p>
3837
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3838<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3839<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3840 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3841 values. Both arguments must have identical types.</p>
3842
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3843<h5>Semantics:</h5>
3844<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3845 This instruction always performs an unsigned division to get the
3846 remainder.</p>
3847
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3848<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3849 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3850
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3851<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3852
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3853<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3854<pre>
3855 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3856</pre>
3857
3858</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3859
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3860<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3861<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3862 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3863</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3864
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3865<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3866
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3867<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3868<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3869 &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]3870</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3871
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3872<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3873<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3874 division of its two operands. This instruction can also take
3875 <a href="#t_vector">vector</a> versions of the values in which case the
3876 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]3877
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3878<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3879<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3880 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3881 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3882
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3883<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3884<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +0000[diff] [blame]3885 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3886 <i>modulo</i> operator (where the result is either zero or has the same sign
3887 as the divisor, <tt>op2</tt>) of a value.
3888 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3889 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3890 Math Forum</a>. For a table of how this is implemented in various languages,
3891 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3892 Wikipedia: modulo operation</a>.</p>
3893
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3894<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3895 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3896
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3897<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3898 Overflow also leads to undefined behavior; this is a rare case, but can
3899 occur, for example, by taking the remainder of a 32-bit division of
3900 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3901 lets srem be implemented using instructions that return both the result of
3902 the division and the remainder.)</p>
3903
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3904<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3905<pre>
3906 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3907</pre>
3908
3909</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3910
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3911<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3912<h4>
3913 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3914</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3915
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3916<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3917
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3918<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3919<pre>
3920 &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]3921</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3922
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3923<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3924<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3925 its two operands.</p>
3926
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3927<h5>Arguments:</h5>
3928<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3929 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3930 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3931
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3932<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3933<p>This instruction returns the <i>remainder</i> of a division. The remainder
3934 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3935
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3936<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3937<pre>
3938 &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]3939</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3940
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3941</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]3942
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3943</div>
3944
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]3945<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3946<h3>
3947 <a name="bitwiseops">Bitwise Binary Operations</a>
3948</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3949
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3950<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3951
3952<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3953 program. They are generally very efficient instructions and can commonly be
3954 strength reduced from other instructions. They require two operands of the
3955 same type, execute an operation on them, and produce a single value. The
3956 resulting value is the same type as its operands.</p>
3957
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3958<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3959<h4>
3960 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3961</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3962
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3963<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3964
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3965<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3966<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3967 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3968 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3969 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3970 &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]3971</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3972
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3973<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3974<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3975 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3976
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3977<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3978<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3979 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3980 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3981
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3982<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3983<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3984 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3985 is (statically or dynamically) negative or equal to or larger than the number
3986 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3987 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3988 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3989
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3990<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3991 <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]3992 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3993 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3994 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3995 they would if the shift were expressed as a mul instruction with the same
3996 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3997
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3998<h5>Example:</h5>
3999<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4000 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
4001 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
4002 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4003 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4004 &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]4005</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4006
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4007</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4008
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4009<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4010<h4>
4011 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
4012</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4013
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4014<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4015
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4016<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4017<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4018 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4019 &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]4020</pre>
4021
4022<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4023<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
4024 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4025
4026<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4027<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4028 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4029 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4030
4031<h5>Semantics:</h5>
4032<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4033 significant bits of the result will be filled with zero bits after the shift.
4034 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4035 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4036 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4037 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4038
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4039<p>If the <tt>exact</tt> keyword is present, the result value of the
4040 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
4041 shifted out are non-zero.</p>
4042
4043
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4044<h5>Example:</h5>
4045<pre>
4046 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4047 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4048 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4049 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4050 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4051 &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]4052</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4053
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4054</div>
4055
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4056<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4057<h4>
4058 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4059</h4>
4060
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4061<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4062
4063<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4064<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4065 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4066 &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]4067</pre>
4068
4069<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4070<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4071 operand shifted to the right a specified number of bits with sign
4072 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4073
4074<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4075<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4076 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4077 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4078
4079<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4080<p>This instruction always performs an arithmetic shift right operation, The
4081 most significant bits of the result will be filled with the sign bit
4082 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4083 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4084 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4085 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4086
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4087<p>If the <tt>exact</tt> keyword is present, the result value of the
4088 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
4089 shifted out are non-zero.</p>
4090
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4091<h5>Example:</h5>
4092<pre>
4093 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4094 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4095 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4096 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4097 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4098 &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]4099</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4100
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4101</div>
4102
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4103<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4104<h4>
4105 <a name="i_and">'<tt>and</tt>' Instruction</a>
4106</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4107
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4108<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4109
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4110<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4111<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4112 &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]4113</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4114
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4115<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4116<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4117 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4118
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4119<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4120<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4121 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4122 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4123
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4124<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4125<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4126
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4127<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4128 <tbody>
4129 <tr>
4130 <td>In0</td>
4131 <td>In1</td>
4132 <td>Out</td>
4133 </tr>
4134 <tr>
4135 <td>0</td>
4136 <td>0</td>
4137 <td>0</td>
4138 </tr>
4139 <tr>
4140 <td>0</td>
4141 <td>1</td>
4142 <td>0</td>
4143 </tr>
4144 <tr>
4145 <td>1</td>
4146 <td>0</td>
4147 <td>0</td>
4148 </tr>
4149 <tr>
4150 <td>1</td>
4151 <td>1</td>
4152 <td>1</td>
4153 </tr>
4154 </tbody>
4155</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4156
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4157<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4158<pre>
4159 &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]4160 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4161 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4162</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4163</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4164<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4165<h4>
4166 <a name="i_or">'<tt>or</tt>' Instruction</a>
4167</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4168
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4169<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4170
4171<h5>Syntax:</h5>
4172<pre>
4173 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4174</pre>
4175
4176<h5>Overview:</h5>
4177<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4178 two operands.</p>
4179
4180<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4181<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4182 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4183 values. Both arguments must have identical types.</p>
4184
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4185<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4186<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4187
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4188<table border="1" cellspacing="0" cellpadding="4">
4189 <tbody>
4190 <tr>
4191 <td>In0</td>
4192 <td>In1</td>
4193 <td>Out</td>
4194 </tr>
4195 <tr>
4196 <td>0</td>
4197 <td>0</td>
4198 <td>0</td>
4199 </tr>
4200 <tr>
4201 <td>0</td>
4202 <td>1</td>
4203 <td>1</td>
4204 </tr>
4205 <tr>
4206 <td>1</td>
4207 <td>0</td>
4208 <td>1</td>
4209 </tr>
4210 <tr>
4211 <td>1</td>
4212 <td>1</td>
4213 <td>1</td>
4214 </tr>
4215 </tbody>
4216</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4217
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4218<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4219<pre>
4220 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4221 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4222 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4223</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4224
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4225</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4226
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4227<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4228<h4>
4229 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4230</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4231
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4232<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4233
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4234<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4235<pre>
4236 &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]4237</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4238
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4239<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4240<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4241 its two operands. The <tt>xor</tt> is used to implement the "one's
4242 complement" operation, which is the "~" operator in C.</p>
4243
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4244<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4245<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4246 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4247 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4248
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4249<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4250<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4251
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4252<table border="1" cellspacing="0" cellpadding="4">
4253 <tbody>
4254 <tr>
4255 <td>In0</td>
4256 <td>In1</td>
4257 <td>Out</td>
4258 </tr>
4259 <tr>
4260 <td>0</td>
4261 <td>0</td>
4262 <td>0</td>
4263 </tr>
4264 <tr>
4265 <td>0</td>
4266 <td>1</td>
4267 <td>1</td>
4268 </tr>
4269 <tr>
4270 <td>1</td>
4271 <td>0</td>
4272 <td>1</td>
4273 </tr>
4274 <tr>
4275 <td>1</td>
4276 <td>1</td>
4277 <td>0</td>
4278 </tr>
4279 </tbody>
4280</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4281
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4282<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4283<pre>
4284 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4285 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4286 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4287 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4288</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4289
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4290</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4291
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4292</div>
4293
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4294<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4295<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4296 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4297</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4298
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4299<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4300
4301<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4302 target-independent manner. These instructions cover the element-access and
4303 vector-specific operations needed to process vectors effectively. While LLVM
4304 does directly support these vector operations, many sophisticated algorithms
4305 will want to use target-specific intrinsics to take full advantage of a
4306 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4307
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4308<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4309<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4310 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4311</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4312
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4313<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4314
4315<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4316<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4317 &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]4318</pre>
4319
4320<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4321<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4322 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4323
4324
4325<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4326<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4327 of <a href="#t_vector">vector</a> type. The second operand is an index
4328 indicating the position from which to extract the element. The index may be
4329 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4330
4331<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4332<p>The result is a scalar of the same type as the element type of
4333 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4334 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4335 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4336
4337<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4338<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4339 &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]4340</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4341
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4342</div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4343
4344<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4345<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4346 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4347</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4348
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4349<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4350
4351<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4352<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +0000[diff] [blame]4353 &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]4354</pre>
4355
4356<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4357<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4358 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4359
4360<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4361<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4362 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4363 whose type must equal the element type of the first operand. The third
4364 operand is an index indicating the position at which to insert the value.
4365 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4366
4367<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4368<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4369 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4370 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4371 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4372
4373<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4374<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4375 &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]4376</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4377
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4378</div>
4379
4380<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4381<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4382 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4383</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4384
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4385<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4386
4387<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4388<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4389 &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]4390</pre>
4391
4392<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4393<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4394 from two input vectors, returning a vector with the same element type as the
4395 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4396
4397<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4398<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4399 with types that match each other. The third argument is a shuffle mask whose
4400 element type is always 'i32'. The result of the instruction is a vector
4401 whose length is the same as the shuffle mask and whose element type is the
4402 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4403
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4404<p>The shuffle mask operand is required to be a constant vector with either
4405 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4406
4407<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4408<p>The elements of the two input vectors are numbered from left to right across
4409 both of the vectors. The shuffle mask operand specifies, for each element of
4410 the result vector, which element of the two input vectors the result element
4411 gets. The element selector may be undef (meaning "don't care") and the
4412 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4413
4414<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4415<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4416 &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]4417 &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]4418 &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]4419 &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]4420 &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]4421 &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]4422 &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]4423 &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]4424</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4425
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4426</div>
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]4427
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4428</div>
4429
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4430<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4431<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4432 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4433</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4434
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4435<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4436
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4437<p>LLVM supports several instructions for working with
4438 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4439
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4440<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4441<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4442 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4443</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4444
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4445<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4446
4447<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4448<pre>
4449 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4450</pre>
4451
4452<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4453<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4454 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4455
4456<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4457<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4458 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4459 <a href="#t_array">array</a> type. The operands are constant indices to
4460 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4461 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4462 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4463 <ul>
4464 <li>Since the value being indexed is not a pointer, the first index is
4465 omitted and assumed to be zero.</li>
4466 <li>At least one index must be specified.</li>
4467 <li>Not only struct indices but also array indices must be in
4468 bounds.</li>
4469 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4470
4471<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4472<p>The result is the value at the position in the aggregate specified by the
4473 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4474
4475<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4476<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4477 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4478</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4479
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4480</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4481
4482<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4483<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4484 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4485</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4486
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4487<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4488
4489<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4490<pre>
Bill Wendling194229e2011-07-26 20:42:28 +0000[diff] [blame]4491 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt;{, &lt;idx&gt;}* <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4492</pre>
4493
4494<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4495<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4496 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4497
4498<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4499<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4500 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4501 <a href="#t_array">array</a> type. The second operand is a first-class
4502 value to insert. The following operands are constant indices indicating
4503 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]4504 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4505 value to insert must have the same type as the value identified by the
4506 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4507
4508<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4509<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4510 that of <tt>val</tt> except that the value at the position specified by the
4511 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4512
4513<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4514<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +0000[diff] [blame]4515 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4516 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4517 %agg3 = insertvalue {i32, {float}} %agg1, float %val, 1, 0 <i>; yields {i32 1, float %val}</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4518</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4519
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4520</div>
4521
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4522</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4523
4524<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4525<h3>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4526 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4527</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4528
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4529<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4530
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4531<p>A key design point of an SSA-based representation is how it represents
4532 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]4533 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4534 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4535
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4536<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4537<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4538 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4539</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4540
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4541<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4542
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4543<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4544<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +0000[diff] [blame]4545 &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]4546</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4547
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4548<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4549<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4550 currently executing function, to be automatically released when this function
4551 returns to its caller. The object is always allocated in the generic address
4552 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4553
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4554<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4555<p>The '<tt>alloca</tt>' instruction
4556 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4557 runtime stack, returning a pointer of the appropriate type to the program.
4558 If "NumElements" is specified, it is the number of elements allocated,
4559 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4560 specified, the value result of the allocation is guaranteed to be aligned to
4561 at least that boundary. If not specified, or if zero, the target can choose
4562 to align the allocation on any convenient boundary compatible with the
4563 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4564
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4565<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4566
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4567<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +0000[diff] [blame]4568<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4569 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4570 memory is automatically released when the function returns. The
4571 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4572 variables that must have an address available. When the function returns
4573 (either with the <tt><a href="#i_ret">ret</a></tt>
4574 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4575 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4576
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4577<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4578<pre>
Dan Gohman81e21672009-01-04 23:49:44 +0000[diff] [blame]4579 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4580 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4581 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4582 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4583</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4584
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4585</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4586
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4587<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4588<h4>
4589 <a name="i_load">'<tt>load</tt>' Instruction</a>
4590</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4591
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4592<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4593
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4594<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4595<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]4596 &lt;result&gt; = load [volatile] &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4597 &lt;result&gt; = load atomic [volatile] &lt;ty&gt;* &lt;pointer&gt; [singlethread] &lt;ordering&gt;, align &lt;alignment&gt;
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4598 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4599</pre>
4600
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4601<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4602<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4603
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4604<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4605<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4606 from which to load. The pointer must point to
4607 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4608 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4609 number or order of execution of this <tt>load</tt> with other <a
4610 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4611
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]4612<p>If the <code>load</code> is marked as <code>atomic</code>, it takes an extra
4613 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4614 argument. The <code>release</code> and <code>acq_rel</code> orderings are
4615 not valid on <code>load</code> instructions. Atomic loads produce <a
4616 href="#memorymodel">defined</a> results when they may see multiple atomic
4617 stores. The type of the pointee must be an integer type whose bit width
4618 is a power of two greater than or equal to eight and less than or equal
4619 to a target-specific size limit. <code>align</code> must be explicitly
4620 specified on atomic loads, and the load has undefined behavior if the
4621 alignment is not set to a value which is at least the size in bytes of
4622 the pointee. <code>!nontemporal</code> does not have any defined semantics
4623 for atomic loads.</p>
4624
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4625<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4626 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]4627 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4628 alignment for the target. It is the responsibility of the code emitter to
4629 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4630 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4631 produce less efficient code. An alignment of 1 is always safe.</p>
4632
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4633<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4634 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4635 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4636 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4637 and code generator that this load is not expected to be reused in the cache.
4638 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4639 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4640
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4641<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4642<p>The location of memory pointed to is loaded. If the value being loaded is of
4643 scalar type then the number of bytes read does not exceed the minimum number
4644 of bytes needed to hold all bits of the type. For example, loading an
4645 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4646 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4647 is undefined if the value was not originally written using a store of the
4648 same type.</p>
4649
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4650<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4651<pre>
4652 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4653 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4654 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4655</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4656
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4657</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4658
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4659<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4660<h4>
4661 <a name="i_store">'<tt>store</tt>' Instruction</a>
4662</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4663
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4664<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4665
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4666<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4667<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]4668 store [volatile] &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
4669 store atomic [volatile] &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt; [singlethread] &lt;ordering&gt;, align &lt;alignment&gt; <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4670</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4671
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4672<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4673<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4674
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4675<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4676<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4677 and an address at which to store it. The type of the
4678 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4679 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4680 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4681 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4682 order of execution of this <tt>store</tt> with other <a
4683 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4684
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]4685<p>If the <code>store</code> is marked as <code>atomic</code>, it takes an extra
4686 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4687 argument. The <code>acquire</code> and <code>acq_rel</code> orderings aren't
4688 valid on <code>store</code> instructions. Atomic loads produce <a
4689 href="#memorymodel">defined</a> results when they may see multiple atomic
4690 stores. The type of the pointee must be an integer type whose bit width
4691 is a power of two greater than or equal to eight and less than or equal
4692 to a target-specific size limit. <code>align</code> must be explicitly
4693 specified on atomic stores, and the store has undefined behavior if the
4694 alignment is not set to a value which is at least the size in bytes of
4695 the pointee. <code>!nontemporal</code> does not have any defined semantics
4696 for atomic stores.</p>
4697
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4698<p>The optional constant "align" argument specifies the alignment of the
4699 operation (that is, the alignment of the memory address). A value of 0 or an
4700 omitted "align" argument means that the operation has the preferential
4701 alignment for the target. It is the responsibility of the code emitter to
4702 ensure that the alignment information is correct. Overestimating the
4703 alignment results in an undefined behavior. Underestimating the alignment may
4704 produce less efficient code. An alignment of 1 is always safe.</p>
4705
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4706<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4707 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4708 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4709 instruction tells the optimizer and code generator that this load is
4710 not expected to be reused in the cache. The code generator may
4711 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4712 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4713
4714
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4715<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4716<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4717 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4718 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4719 does not exceed the minimum number of bytes needed to hold all bits of the
4720 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4721 writing a value of a type like <tt>i20</tt> with a size that is not an
4722 integral number of bytes, it is unspecified what happens to the extra bits
4723 that do not belong to the type, but they will typically be overwritten.</p>
4724
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4725<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4726<pre>
4727 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]4728 store i32 3, i32* %ptr <i>; yields {void}</i>
4729 %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]4730</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4731
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]4732</div>
4733
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4734<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4735<h4>
4736<a name="i_fence">'<tt>fence</tt>' Instruction</a>
4737</h4>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]4738
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4739<div>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]4740
4741<h5>Syntax:</h5>
4742<pre>
4743 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
4744</pre>
4745
4746<h5>Overview:</h5>
4747<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
4748between operations.</p>
4749
4750<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
4751href="#ordering">ordering</a> argument which defines what
4752<i>synchronizes-with</i> edges they add. They can only be given
4753<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
4754<code>seq_cst</code> orderings.</p>
4755
4756<h5>Semantics:</h5>
4757<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
4758semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
4759<code>acquire</code> ordering semantics if and only if there exist atomic
4760operations <var>X</var> and <var>Y</var>, both operating on some atomic object
4761<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
4762<var>X</var> modifies <var>M</var> (either directly or through some side effect
4763of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
4764<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
4765<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
4766than an explicit <code>fence</code>, one (but not both) of the atomic operations
4767<var>X</var> or <var>Y</var> might provide a <code>release</code> or
4768<code>acquire</code> (resp.) ordering constraint and still
4769<i>synchronize-with</i> the explicit <code>fence</code> and establish the
4770<i>happens-before</i> edge.</p>
4771
4772<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
4773having both <code>acquire</code> and <code>release</code> semantics specified
4774above, participates in the global program order of other <code>seq_cst</code>
4775operations and/or fences.</p>
4776
4777<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
4778specifies that the fence only synchronizes with other fences in the same
4779thread. (This is useful for interacting with signal handlers.)</p>
4780
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]4781<h5>Example:</h5>
4782<pre>
4783 fence acquire <i>; yields {void}</i>
4784 fence singlethread seq_cst <i>; yields {void}</i>
4785</pre>
4786
4787</div>
4788
4789<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4790<h4>
4791<a name="i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a>
4792</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4793
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4794<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4795
4796<h5>Syntax:</h5>
4797<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]4798 cmpxchg [volatile] &lt;ty&gt;* &lt;pointer&gt;, &lt;ty&gt; &lt;cmp&gt;, &lt;ty&gt; &lt;new&gt; [singlethread] &lt;ordering&gt; <i>; yields {ty}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4799</pre>
4800
4801<h5>Overview:</h5>
4802<p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
4803It loads a value in memory and compares it to a given value. If they are
4804equal, it stores a new value into the memory.</p>
4805
4806<h5>Arguments:</h5>
4807<p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
4808address to operate on, a value to compare to the value currently be at that
4809address, and a new value to place at that address if the compared values are
4810equal. The type of '<var>&lt;cmp&gt;</var>' must be an integer type whose
4811bit width is a power of two greater than or equal to eight and less than
4812or equal to a target-specific size limit. '<var>&lt;cmp&gt;</var>' and
4813'<var>&lt;new&gt;</var>' must have the same type, and the type of
4814'<var>&lt;pointer&gt;</var>' must be a pointer to that type. If the
4815<code>cmpxchg</code> is marked as <code>volatile</code>, then the
4816optimizer is not allowed to modify the number or order of execution
4817of this <code>cmpxchg</code> with other <a href="#volatile">volatile
4818operations</a>.</p>
4819
4820<!-- FIXME: Extend allowed types. -->
4821
4822<p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
4823<code>cmpxchg</code> synchronizes with other atomic operations.</p>
4824
4825<p>The optional "<code>singlethread</code>" argument declares that the
4826<code>cmpxchg</code> is only atomic with respect to code (usually signal
4827handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
4828cmpxchg is atomic with respect to all other code in the system.</p>
4829
4830<p>The pointer passed into cmpxchg must have alignment greater than or equal to
4831the size in memory of the operand.
4832
4833<h5>Semantics:</h5>
4834<p>The contents of memory at the location specified by the
4835'<tt>&lt;pointer&gt;</tt>' operand is read and compared to
4836'<tt>&lt;cmp&gt;</tt>'; if the read value is the equal,
4837'<tt>&lt;new&gt;</tt>' is written. The original value at the location
4838is returned.
4839
4840<p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
4841purpose of identifying <a href="#release_sequence">release sequences</a>. A
4842failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
4843parameter determined by dropping any <code>release</code> part of the
4844<code>cmpxchg</code>'s ordering.</p>
4845
4846<!--
4847FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
4848optimization work on ARM.)
4849
4850FIXME: Is a weaker ordering constraint on failure helpful in practice?
4851-->
4852
4853<h5>Example:</h5>
4854<pre>
4855entry:
4856 %orig = atomic <a href="#i_load">load</a> i32* %ptr unordered <i>; yields {i32}</i>
4857 <a href="#i_br">br</a> label %loop
4858
4859loop:
4860 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
4861 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
4862 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
4863 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
4864 <a href="#i_br">br</a> i1 %success, label %done, label %loop
4865
4866done:
4867 ...
4868</pre>
4869
4870</div>
4871
4872<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4873<h4>
4874<a name="i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a>
4875</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4876
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4877<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4878
4879<h5>Syntax:</h5>
4880<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]4881 atomicrmw [volatile] &lt;operation&gt; &lt;ty&gt;* &lt;pointer&gt;, &lt;ty&gt; &lt;value&gt; [singlethread] &lt;ordering&gt; <i>; yields {ty}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4882</pre>
4883
4884<h5>Overview:</h5>
4885<p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
4886
4887<h5>Arguments:</h5>
4888<p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
4889operation to apply, an address whose value to modify, an argument to the
4890operation. The operation must be one of the following keywords:</p>
4891<ul>
4892 <li>xchg</li>
4893 <li>add</li>
4894 <li>sub</li>
4895 <li>and</li>
4896 <li>nand</li>
4897 <li>or</li>
4898 <li>xor</li>
4899 <li>max</li>
4900 <li>min</li>
4901 <li>umax</li>
4902 <li>umin</li>
4903</ul>
4904
4905<p>The type of '<var>&lt;value&gt;</var>' must be an integer type whose
4906bit width is a power of two greater than or equal to eight and less than
4907or equal to a target-specific size limit. The type of the
4908'<code>&lt;pointer&gt;</code>' operand must be a pointer to that type.
4909If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
4910optimizer is not allowed to modify the number or order of execution of this
4911<code>atomicrmw</code> with other <a href="#volatile">volatile
4912 operations</a>.</p>
4913
4914<!-- FIXME: Extend allowed types. -->
4915
4916<h5>Semantics:</h5>
4917<p>The contents of memory at the location specified by the
4918'<tt>&lt;pointer&gt;</tt>' operand are atomically read, modified, and written
4919back. The original value at the location is returned. The modification is
4920specified by the <var>operation</var> argument:</p>
4921
4922<ul>
4923 <li>xchg: <code>*ptr = val</code></li>
4924 <li>add: <code>*ptr = *ptr + val</code></li>
4925 <li>sub: <code>*ptr = *ptr - val</code></li>
4926 <li>and: <code>*ptr = *ptr &amp; val</code></li>
4927 <li>nand: <code>*ptr = ~(*ptr &amp; val)</code></li>
4928 <li>or: <code>*ptr = *ptr | val</code></li>
4929 <li>xor: <code>*ptr = *ptr ^ val</code></li>
4930 <li>max: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using a signed comparison)</li>
4931 <li>min: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using a signed comparison)</li>
4932 <li>umax: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using an unsigned comparison)</li>
4933 <li>umin: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using an unsigned comparison)</li>
4934</ul>
4935
4936<h5>Example:</h5>
4937<pre>
4938 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
4939</pre>
4940
4941</div>
4942
4943<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4944<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4945 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4946</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4947
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4948<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4949
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4950<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4951<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4952 &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]4953 &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]4954</pre>
4955
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4956<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4957<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4958 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4959 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4960
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4961<h5>Arguments:</h5>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4962<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]4963 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4964 elements of the aggregate object are indexed. The interpretation of each
4965 index is dependent on the type being indexed into. The first index always
4966 indexes the pointer value given as the first argument, the second index
4967 indexes a value of the type pointed to (not necessarily the value directly
4968 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4969 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4970 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4971 can never be pointers, since that would require loading the pointer before
4972 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4973
4974<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]4975 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4976 integer <b>constants</b> are allowed. When indexing into an array, pointer
4977 or vector, integers of any width are allowed, and they are not required to be
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]4978 constant. These integers are treated as signed values where relevant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4979
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4980<p>For example, let's consider a C code fragment and how it gets compiled to
4981 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4982
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4983<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4984struct RT {
4985 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4986 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4987 char C;
4988};
4989struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4990 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4991 double Y;
4992 struct RT Z;
4993};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4994
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4995int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4996 return &amp;s[1].Z.B[5][13];
4997}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4998</pre>
4999
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5000<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5001
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5002<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]5003%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
5004%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5005
Dan Gohman4df605b2009-07-25 02:23:48 +0000[diff] [blame]5006define i32* @foo(%ST* %s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5007entry:
5008 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
5009 ret i32* %reg
5010}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5011</pre>
5012
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5013<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5014<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]5015 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
5016 }</tt>' type, a structure. The second index indexes into the third element
5017 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
5018 i8 }</tt>' type, another structure. The third index indexes into the second
5019 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
5020 array. The two dimensions of the array are subscripted into, yielding an
5021 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
5022 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5023
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5024<p>Note that it is perfectly legal to index partially through a structure,
5025 returning a pointer to an inner element. Because of this, the LLVM code for
5026 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5027
5028<pre>
Dan Gohman4df605b2009-07-25 02:23:48 +0000[diff] [blame]5029 define i32* @foo(%ST* %s) {
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5030 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]5031 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
5032 %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]5033 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
5034 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
5035 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5036 }
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5037</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]5038
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5039<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]5040 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
5041 base pointer is not an <i>in bounds</i> address of an allocated object,
5042 or if any of the addresses that would be formed by successive addition of
5043 the offsets implied by the indices to the base address with infinitely
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5044 precise signed arithmetic are not an <i>in bounds</i> address of that
5045 allocated object. The <i>in bounds</i> addresses for an allocated object
5046 are all the addresses that point into the object, plus the address one
5047 byte past the end.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5048
5049<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5050 the base address with silently-wrapping two's complement arithmetic. If the
5051 offsets have a different width from the pointer, they are sign-extended or
5052 truncated to the width of the pointer. The result value of the
5053 <tt>getelementptr</tt> may be outside the object pointed to by the base
5054 pointer. The result value may not necessarily be used to access memory
5055 though, even if it happens to point into allocated storage. See the
5056 <a href="#pointeraliasing">Pointer Aliasing Rules</a> section for more
5057 information.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5058
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5059<p>The getelementptr instruction is often confusing. For some more insight into
5060 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]5061
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5062<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5063<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5064 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5065 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5066 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5067 %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]5068 <i>; yields i8*:eptr</i>
5069 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +0000[diff] [blame]5070 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +0000[diff] [blame]5071 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5072</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5073
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5074</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5075
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5076</div>
5077
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5078<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5079<h3>
5080 <a name="convertops">Conversion Operations</a>
5081</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5082
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5083<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5084
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5085<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5086 which all take a single operand and a type. They perform various bit
5087 conversions on the operand.</p>
5088
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5089<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5090<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5091 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5092</h4>
5093
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5094<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5095
5096<h5>Syntax:</h5>
5097<pre>
5098 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5099</pre>
5100
5101<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5102<p>The '<tt>trunc</tt>' instruction truncates its operand to the
5103 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5104
5105<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5106<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5107 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5108 of the same number of integers.
5109 The bit size of the <tt>value</tt> must be larger than
5110 the bit size of the destination type, <tt>ty2</tt>.
5111 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5112
5113<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5114<p>The '<tt>trunc</tt>' instruction truncates the high order bits
5115 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5116 source size must be larger than the destination size, <tt>trunc</tt> cannot
5117 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5118
5119<h5>Example:</h5>
5120<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5121 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5122 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5123 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5124 %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]5125</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5126
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5127</div>
5128
5129<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5130<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5131 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5132</h4>
5133
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5134<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5135
5136<h5>Syntax:</h5>
5137<pre>
5138 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5139</pre>
5140
5141<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5142<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5143 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5144
5145
5146<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5147<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5148 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5149 of the same number of integers.
5150 The bit size of the <tt>value</tt> must be smaller than
5151 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5152 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5153
5154<h5>Semantics:</h5>
5155<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]5156 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5157
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]5158<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]5159
5160<h5>Example:</h5>
5161<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5162 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5163 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5164 %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]5165</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5166
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5167</div>
5168
5169<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5170<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5171 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5172</h4>
5173
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5174<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5175
5176<h5>Syntax:</h5>
5177<pre>
5178 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5179</pre>
5180
5181<h5>Overview:</h5>
5182<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5183
5184<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5185<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5186 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5187 of the same number of integers.
5188 The bit size of the <tt>value</tt> must be smaller than
5189 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5190 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5191
5192<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5193<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5194 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5195 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5196
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5197<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]5198
5199<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5200<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5201 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5202 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5203 %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]5204</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5205
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5206</div>
5207
5208<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5209<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5210 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5211</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5212
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5213<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5214
5215<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5216<pre>
5217 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5218</pre>
5219
5220<h5>Overview:</h5>
5221<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5222 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5223
5224<h5>Arguments:</h5>
5225<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5226 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5227 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]5228 <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]5229 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5230
5231<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5232<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5233 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5234 <a href="#t_floating">floating point</a> type. If the value cannot fit
5235 within the destination type, <tt>ty2</tt>, then the results are
5236 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5237
5238<h5>Example:</h5>
5239<pre>
5240 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5241 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5242</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5243
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5244</div>
5245
5246<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5247<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5248 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5249</h4>
5250
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5251<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5252
5253<h5>Syntax:</h5>
5254<pre>
5255 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5256</pre>
5257
5258<h5>Overview:</h5>
5259<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]5260 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5261
5262<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5263<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5264 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5265 a <a href="#t_floating">floating point</a> type to cast it to. The source
5266 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5267
5268<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5269<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5270 <a href="#t_floating">floating point</a> type to a larger
5271 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5272 used to make a <i>no-op cast</i> because it always changes bits. Use
5273 <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]5274
5275<h5>Example:</h5>
5276<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +0000[diff] [blame]5277 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5278 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5279</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5280
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5281</div>
5282
5283<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5284<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]5285 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5286</h4>
5287
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5288<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5289
5290<h5>Syntax:</h5>
5291<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5292 &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]5293</pre>
5294
5295<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5296<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5297 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5298
5299<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5300<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5301 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5302 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5303 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5304 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5305
5306<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5307<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5308 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5309 towards zero) unsigned integer value. If the value cannot fit
5310 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5311
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5312<h5>Example:</h5>
5313<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5314 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5315 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5316 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5317</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5318
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5319</div>
5320
5321<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5322<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5323 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5324</h4>
5325
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5326<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5327
5328<h5>Syntax:</h5>
5329<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5330 &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]5331</pre>
5332
5333<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5334<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5335 <a href="#t_floating">floating point</a> <tt>value</tt> to
5336 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5337
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5338<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5339<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5340 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5341 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5342 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5343 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5344
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5345<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5346<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5347 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5348 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5349 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5350
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5351<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5352<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5353 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5354 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5355 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5356</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5357
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5358</div>
5359
5360<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5361<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5362 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5363</h4>
5364
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5365<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5366
5367<h5>Syntax:</h5>
5368<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5369 &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]5370</pre>
5371
5372<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5373<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5374 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5375
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5376<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5377<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]5378 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5379 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5380 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5381 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5382
5383<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5384<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5385 integer quantity and converts it to the corresponding floating point
5386 value. If the value cannot fit in the floating point value, the results are
5387 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5388
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5389<h5>Example:</h5>
5390<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5391 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5392 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5393</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5394
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5395</div>
5396
5397<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5398<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5399 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5400</h4>
5401
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5402<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5403
5404<h5>Syntax:</h5>
5405<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5406 &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]5407</pre>
5408
5409<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5410<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5411 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5412
5413<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5414<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]5415 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5416 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5417 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5418 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5419
5420<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5421<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5422 quantity and converts it to the corresponding floating point value. If the
5423 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5424
5425<h5>Example:</h5>
5426<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5427 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5428 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5429</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5430
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5431</div>
5432
5433<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5434<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5435 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5436</h4>
5437
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5438<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5439
5440<h5>Syntax:</h5>
5441<pre>
5442 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5443</pre>
5444
5445<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5446<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5447 the integer type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5448
5449<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5450<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5451 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5452 <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]5453
5454<h5>Semantics:</h5>
5455<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5456 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5457 truncating or zero extending that value to the size of the integer type. If
5458 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5459 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5460 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5461 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5462
5463<h5>Example:</h5>
5464<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5465 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5466 %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]5467</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5468
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5469</div>
5470
5471<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5472<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5473 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5474</h4>
5475
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5476<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5477
5478<h5>Syntax:</h5>
5479<pre>
5480 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5481</pre>
5482
5483<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5484<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5485 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5486
5487<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]5488<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5489 value to cast, and a type to cast it to, which must be a
5490 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5491
5492<h5>Semantics:</h5>
5493<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5494 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5495 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5496 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5497 than the size of a pointer then a zero extension is done. If they are the
5498 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5499
5500<h5>Example:</h5>
5501<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5502 %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]5503 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5504 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5505</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5506
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5507</div>
5508
5509<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5510<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5511 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5512</h4>
5513
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5514<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5515
5516<h5>Syntax:</h5>
5517<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5518 &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]5519</pre>
5520
5521<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5522<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5523 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5524
5525<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5526<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5527 non-aggregate first class value, and a type to cast it to, which must also be
5528 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5529 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5530 identical. If the source type is a pointer, the destination type must also be
5531 a pointer. This instruction supports bitwise conversion of vectors to
5532 integers and to vectors of other types (as long as they have the same
5533 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5534
5535<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5536<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5537 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5538 this conversion. The conversion is done as if the <tt>value</tt> had been
5539 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5540 be converted to other pointer types with this instruction. To convert
5541 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5542 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5543
5544<h5>Example:</h5>
5545<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5546 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5547 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5548 %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]5549</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5550
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5551</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5552
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5553</div>
5554
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5555<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5556<h3>
5557 <a name="otherops">Other Operations</a>
5558</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5559
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5560<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5561
5562<p>The instructions in this category are the "miscellaneous" instructions, which
5563 defy better classification.</p>
5564
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5565<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5566<h4>
5567 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5568</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5569
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5570<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5571
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5572<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5573<pre>
5574 &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]5575</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5576
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5577<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5578<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5579 boolean values based on comparison of its two integer, integer vector, or
5580 pointer operands.</p>
5581
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5582<h5>Arguments:</h5>
5583<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5584 the condition code indicating the kind of comparison to perform. It is not a
5585 value, just a keyword. The possible condition code are:</p>
5586
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5587<ol>
5588 <li><tt>eq</tt>: equal</li>
5589 <li><tt>ne</tt>: not equal </li>
5590 <li><tt>ugt</tt>: unsigned greater than</li>
5591 <li><tt>uge</tt>: unsigned greater or equal</li>
5592 <li><tt>ult</tt>: unsigned less than</li>
5593 <li><tt>ule</tt>: unsigned less or equal</li>
5594 <li><tt>sgt</tt>: signed greater than</li>
5595 <li><tt>sge</tt>: signed greater or equal</li>
5596 <li><tt>slt</tt>: signed less than</li>
5597 <li><tt>sle</tt>: signed less or equal</li>
5598</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5599
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]5600<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5601 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5602 typed. They must also be identical types.</p>
5603
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5604<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5605<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5606 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5607 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]5608 result, as follows:</p>
5609
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5610<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5611 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5612 <tt>false</tt> otherwise. No sign interpretation is necessary or
5613 performed.</li>
5614
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5615 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5616 <tt>false</tt> otherwise. No sign interpretation is necessary or
5617 performed.</li>
5618
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5619 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5620 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5621
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5622 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5623 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5624 to <tt>op2</tt>.</li>
5625
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5626 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5627 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5628
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5629 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5630 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5631
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5632 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5633 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5634
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5635 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5636 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5637 to <tt>op2</tt>.</li>
5638
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5639 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5640 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5641
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5642 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5643 <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]5644</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5645
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5646<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5647 values are compared as if they were integers.</p>
5648
5649<p>If the operands are integer vectors, then they are compared element by
5650 element. The result is an <tt>i1</tt> vector with the same number of elements
5651 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5652
5653<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5654<pre>
5655 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5656 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5657 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5658 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5659 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5660 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5661</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5662
5663<p>Note that the code generator does not yet support vector types with
5664 the <tt>icmp</tt> instruction.</p>
5665
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5666</div>
5667
5668<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5669<h4>
5670 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5671</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5672
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5673<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5674
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5675<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5676<pre>
5677 &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]5678</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5679
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5680<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5681<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5682 values based on comparison of its operands.</p>
5683
5684<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]5685(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5686
5687<p>If the operands are floating point vectors, then the result type is a vector
5688 of boolean with the same number of elements as the operands being
5689 compared.</p>
5690
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5691<h5>Arguments:</h5>
5692<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5693 the condition code indicating the kind of comparison to perform. It is not a
5694 value, just a keyword. The possible condition code are:</p>
5695
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5696<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5697 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5698 <li><tt>oeq</tt>: ordered and equal</li>
5699 <li><tt>ogt</tt>: ordered and greater than </li>
5700 <li><tt>oge</tt>: ordered and greater than or equal</li>
5701 <li><tt>olt</tt>: ordered and less than </li>
5702 <li><tt>ole</tt>: ordered and less than or equal</li>
5703 <li><tt>one</tt>: ordered and not equal</li>
5704 <li><tt>ord</tt>: ordered (no nans)</li>
5705 <li><tt>ueq</tt>: unordered or equal</li>
5706 <li><tt>ugt</tt>: unordered or greater than </li>
5707 <li><tt>uge</tt>: unordered or greater than or equal</li>
5708 <li><tt>ult</tt>: unordered or less than </li>
5709 <li><tt>ule</tt>: unordered or less than or equal</li>
5710 <li><tt>une</tt>: unordered or not equal</li>
5711 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5712 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5713</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5714
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5715<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5716 <i>unordered</i> means that either operand may be a QNAN.</p>
5717
5718<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5719 a <a href="#t_floating">floating point</a> type or
5720 a <a href="#t_vector">vector</a> of floating point type. They must have
5721 identical types.</p>
5722
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5723<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]5724<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5725 according to the condition code given as <tt>cond</tt>. If the operands are
5726 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5727 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5728 follows:</p>
5729
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5730<ol>
5731 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5732
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5733 <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]5734 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5735
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5736 <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]5737 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5738
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5739 <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]5740 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5741
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5742 <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]5743 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5744
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5745 <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]5746 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5747
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5748 <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]5749 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5750
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5751 <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]5752
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5753 <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]5754 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5755
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5756 <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]5757 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5758
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5759 <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]5760 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5761
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5762 <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]5763 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5764
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5765 <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]5766 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5767
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5768 <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]5769 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5770
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5771 <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]5772
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5773 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5774</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5775
5776<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5777<pre>
5778 &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]5779 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5780 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5781 &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]5782</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5783
5784<p>Note that the code generator does not yet support vector types with
5785 the <tt>fcmp</tt> instruction.</p>
5786
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5787</div>
5788
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5789<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5790<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5791 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5792</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5793
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5794<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5795
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5796<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5797<pre>
5798 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5799</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5800
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5801<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5802<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5803 SSA graph representing the function.</p>
5804
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5805<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5806<p>The type of the incoming values is specified with the first type field. After
5807 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5808 one pair for each predecessor basic block of the current block. Only values
5809 of <a href="#t_firstclass">first class</a> type may be used as the value
5810 arguments to the PHI node. Only labels may be used as the label
5811 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5812
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5813<p>There must be no non-phi instructions between the start of a basic block and
5814 the PHI instructions: i.e. PHI instructions must be first in a basic
5815 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5816
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5817<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5818 occur on the edge from the corresponding predecessor block to the current
5819 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5820 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +0000[diff] [blame]5821
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5822<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5823<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5824 specified by the pair corresponding to the predecessor basic block that
5825 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5826
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5827<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5828<pre>
5829Loop: ; Infinite loop that counts from 0 on up...
5830 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5831 %nextindvar = add i32 %indvar, 1
5832 br label %Loop
5833</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5834
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5835</div>
5836
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5837<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5838<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5839 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5840</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5841
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5842<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5843
5844<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5845<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]5846 &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>
5847
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5848 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5849</pre>
5850
5851<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5852<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5853 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5854
5855
5856<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5857<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5858 values indicating the condition, and two values of the
5859 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5860 vectors and the condition is a scalar, then entire vectors are selected, not
5861 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5862
5863<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5864<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5865 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5866
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5867<p>If the condition is a vector of i1, then the value arguments must be vectors
5868 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5869
5870<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5871<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5872 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5873</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5874
5875<p>Note that the code generator does not yet support conditions
5876 with vector type.</p>
5877
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5878</div>
5879
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]5880<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5881<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5882 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5883</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5884
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5885<div>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5886
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5887<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5888<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]5889 &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]5890</pre>
5891
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5892<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5893<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5894
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5895<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5896<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5897
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5898<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5899 <li>The optional "tail" marker indicates that the callee function does not
5900 access any allocas or varargs in the caller. Note that calls may be
5901 marked "tail" even if they do not occur before
5902 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5903 present, the function call is eligible for tail call optimization,
5904 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +0000[diff] [blame]5905 optimized into a jump</a>. The code generator may optimize calls marked
5906 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5907 sibling call optimization</a> when the caller and callee have
5908 matching signatures, or 2) forced tail call optimization when the
5909 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5910 <ul>
5911 <li>Caller and callee both have the calling
5912 convention <tt>fastcc</tt>.</li>
5913 <li>The call is in tail position (ret immediately follows call and ret
5914 uses value of call or is void).</li>
5915 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +0000[diff] [blame]5916 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5917 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5918 constraints are met.</a></li>
5919 </ul>
5920 </li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5921
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5922 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5923 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5924 defaults to using C calling conventions. The calling convention of the
5925 call must match the calling convention of the target function, or else the
5926 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5927
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5928 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5929 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5930 '<tt>inreg</tt>' attributes are valid here.</li>
5931
5932 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5933 type of the return value. Functions that return no value are marked
5934 <tt><a href="#t_void">void</a></tt>.</li>
5935
5936 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5937 being invoked. The argument types must match the types implied by this
5938 signature. This type can be omitted if the function is not varargs and if
5939 the function type does not return a pointer to a function.</li>
5940
5941 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5942 be invoked. In most cases, this is a direct function invocation, but
5943 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5944 to function value.</li>
5945
5946 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]5947 signature argument types and parameter attributes. All arguments must be
5948 of <a href="#t_firstclass">first class</a> type. If the function
5949 signature indicates the function accepts a variable number of arguments,
5950 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5951
5952 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5953 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5954 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5955</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5956
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5957<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5958<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5959 a specified function, with its incoming arguments bound to the specified
5960 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5961 function, control flow continues with the instruction after the function
5962 call, and the return value of the function is bound to the result
5963 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5964
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5965<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5966<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]5967 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]5968 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +0000[diff] [blame]5969 %X = tail call i32 @foo() <i>; yields i32</i>
5970 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5971 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]5972
5973 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5974 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]5975 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5976 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +0000[diff] [blame]5977 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +0000[diff] [blame]5978 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5979</pre>
5980
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]5981<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +0000[diff] [blame]5982standard C99 library as being the C99 library functions, and may perform
5983optimizations or generate code for them under that assumption. This is
5984something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5985freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]5986
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5987</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5988
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5989<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5990<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]5991 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5992</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5993
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5994<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5995
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5996<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5997<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5998 &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]5999</pre>
6000
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6001<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6002<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6003 the "variable argument" area of a function call. It is used to implement the
6004 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6005
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6006<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6007<p>This instruction takes a <tt>va_list*</tt> value and the type of the
6008 argument. It returns a value of the specified argument type and increments
6009 the <tt>va_list</tt> to point to the next argument. The actual type
6010 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6011
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6012<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6013<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
6014 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
6015 to the next argument. For more information, see the variable argument
6016 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6017
6018<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]6019 take a variable number of arguments, for example, the <tt>vfprintf</tt>
6020 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6021
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6022<p><tt>va_arg</tt> is an LLVM instruction instead of
6023 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
6024 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6025
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6026<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6027<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
6028
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6029<p>Note that the code generator does not yet fully support va_arg on many
6030 targets. Also, it does not currently support va_arg with aggregate types on
6031 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]6032
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6033</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6034
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6035<!-- _______________________________________________________________________ -->
6036<h4>
6037 <a name="i_landingpad">'<tt>landingpad</tt>' Instruction</a>
6038</h4>
6039
6040<div>
6041
6042<h5>Syntax:</h5>
6043<pre>
Bill Wendlingbf13ee12011-08-08 08:06:05 +0000[diff] [blame]6044 &lt;resultval&gt; = landingpad &lt;somety&gt; personality &lt;type&gt; &lt;pers_fn&gt; &lt;clause&gt;+
6045 &lt;resultval&gt; = landingpad &lt;somety&gt; personality &lt;type&gt; &lt;pers_fn&gt; cleanup &lt;clause&gt;*
6046
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6047 &lt;clause&gt; := catch &lt;type&gt; &lt;value&gt;
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6048 &lt;clause&gt; := filter &lt;array constant type&gt; &lt;array constant&gt;
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6049</pre>
6050
6051<h5>Overview:</h5>
6052<p>The '<tt>landingpad</tt>' instruction is used by
6053 <a href="ExceptionHandling.html#overview">LLVM's exception handling
6054 system</a> to specify that a basic block is a landing pad &mdash; one where
6055 the exception lands, and corresponds to the code found in the
6056 <i><tt>catch</tt></i> portion of a <i><tt>try/catch</tt></i> sequence. It
6057 defines values supplied by the personality function (<tt>pers_fn</tt>) upon
6058 re-entry to the function. The <tt>resultval</tt> has the
6059 type <tt>somety</tt>.</p>
6060
6061<h5>Arguments:</h5>
6062<p>This instruction takes a <tt>pers_fn</tt> value. This is the personality
6063 function associated with the unwinding mechanism. The optional
6064 <tt>cleanup</tt> flag indicates that the landing pad block is a cleanup.</p>
6065
6066<p>A <tt>clause</tt> begins with the clause type &mdash; <tt>catch</tt>
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6067 or <tt>filter</tt> &mdash; and contains the global variable representing the
6068 "type" that may be caught or filtered respectively. Unlike the
6069 <tt>catch</tt> clause, the <tt>filter</tt> clause takes an array constant as
6070 its argument. Use "<tt>[0 x i8**] undef</tt>" for a filter which cannot
6071 throw. The '<tt>landingpad</tt>' instruction must contain <em>at least</em>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6072 one <tt>clause</tt> or the <tt>cleanup</tt> flag.</p>
6073
6074<h5>Semantics:</h5>
6075<p>The '<tt>landingpad</tt>' instruction defines the values which are set by the
6076 personality function (<tt>pers_fn</tt>) upon re-entry to the function, and
6077 therefore the "result type" of the <tt>landingpad</tt> instruction. As with
6078 calling conventions, how the personality function results are represented in
6079 LLVM IR is target specific.</p>
6080
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6081<p>The clauses are applied in order from top to bottom. If two
6082 <tt>landingpad</tt> instructions are merged together through inlining, the
Bill Wendling2905c322011-08-08 07:58:58 +0000[diff] [blame]6083 clauses from the calling function are appended to the list of clauses.</p>
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6084
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6085<p>The <tt>landingpad</tt> instruction has several restrictions:</p>
6086
6087<ul>
6088 <li>A landing pad block is a basic block which is the unwind destination of an
6089 '<tt>invoke</tt>' instruction.</li>
6090 <li>A landing pad block must have a '<tt>landingpad</tt>' instruction as its
6091 first non-PHI instruction.</li>
6092 <li>There can be only one '<tt>landingpad</tt>' instruction within the landing
6093 pad block.</li>
6094 <li>A basic block that is not a landing pad block may not include a
6095 '<tt>landingpad</tt>' instruction.</li>
6096 <li>All '<tt>landingpad</tt>' instructions in a function must have the same
6097 personality function.</li>
6098</ul>
6099
6100<h5>Example:</h5>
6101<pre>
6102 ;; A landing pad which can catch an integer.
6103 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6104 catch i8** @_ZTIi
6105 ;; A landing pad that is a cleanup.
6106 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6107 cleanup
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6108 ;; A landing pad which can catch an integer and can only throw a double.
6109 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6110 catch i8** @_ZTIi
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6111 filter [1 x i8**] [@_ZTId]
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6112</pre>
6113
6114</div>
6115
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6116</div>
6117
6118</div>
6119
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6120<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6121<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6122<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6123
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6124<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6125
6126<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6127 well known names and semantics and are required to follow certain
6128 restrictions. Overall, these intrinsics represent an extension mechanism for
6129 the LLVM language that does not require changing all of the transformations
6130 in LLVM when adding to the language (or the bitcode reader/writer, the
6131 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6132
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]6133<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6134 prefix is reserved in LLVM for intrinsic names; thus, function names may not
6135 begin with this prefix. Intrinsic functions must always be external
6136 functions: you cannot define the body of intrinsic functions. Intrinsic
6137 functions may only be used in call or invoke instructions: it is illegal to
6138 take the address of an intrinsic function. Additionally, because intrinsic
6139 functions are part of the LLVM language, it is required if any are added that
6140 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6141
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6142<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6143 family of functions that perform the same operation but on different data
6144 types. Because LLVM can represent over 8 million different integer types,
6145 overloading is used commonly to allow an intrinsic function to operate on any
6146 integer type. One or more of the argument types or the result type can be
6147 overloaded to accept any integer type. Argument types may also be defined as
6148 exactly matching a previous argument's type or the result type. This allows
6149 an intrinsic function which accepts multiple arguments, but needs all of them
6150 to be of the same type, to only be overloaded with respect to a single
6151 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6152
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6153<p>Overloaded intrinsics will have the names of its overloaded argument types
6154 encoded into its function name, each preceded by a period. Only those types
6155 which are overloaded result in a name suffix. Arguments whose type is matched
6156 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6157 can take an integer of any width and returns an integer of exactly the same
6158 integer width. This leads to a family of functions such as
6159 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6160 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6161 suffix is required. Because the argument's type is matched against the return
6162 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6163
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6164<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6165 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6166
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6167<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6168<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6169 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6170</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6171
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6172<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6173
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6174<p>Variable argument support is defined in LLVM with
6175 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6176 intrinsic functions. These functions are related to the similarly named
6177 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6178
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6179<p>All of these functions operate on arguments that use a target-specific value
6180 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6181 not define what this type is, so all transformations should be prepared to
6182 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6183
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]6184<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]6185 instruction and the variable argument handling intrinsic functions are
6186 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6187
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]6188<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6189define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6190 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6191 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6192 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6193 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6194
6195 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6196 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6197
6198 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6199 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6200 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6201 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6202 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6203
6204 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6205 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]6206 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6207}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6208
6209declare void @llvm.va_start(i8*)
6210declare void @llvm.va_copy(i8*, i8*)
6211declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6212</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6213
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6214<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6215<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6216 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6217</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6218
6219
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6220<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6221
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6222<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6223<pre>
6224 declare void %llvm.va_start(i8* &lt;arglist&gt;)
6225</pre>
6226
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6227<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6228<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
6229 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]6230
6231<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6232<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]6233
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6234<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6235<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]6236 macro available in C. In a target-dependent way, it initializes
6237 the <tt>va_list</tt> element to which the argument points, so that the next
6238 call to <tt>va_arg</tt> will produce the first variable argument passed to
6239 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6240 need to know the last argument of the function as the compiler can figure
6241 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6242
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6243</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6244
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6245<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6246<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6247 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6248</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6249
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6250<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6251
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6252<h5>Syntax:</h5>
6253<pre>
6254 declare void @llvm.va_end(i8* &lt;arglist&gt;)
6255</pre>
6256
6257<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6258<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6259 which has been initialized previously
6260 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6261 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6262
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6263<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6264<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]6265
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6266<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6267<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]6268 macro available in C. In a target-dependent way, it destroys
6269 the <tt>va_list</tt> element to which the argument points. Calls
6270 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6271 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6272 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6273
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6274</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6275
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6276<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6277<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6278 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6279</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6280
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6281<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6282
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6283<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6284<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6285 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6286</pre>
6287
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6288<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6289<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6290 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6291
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6292<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6293<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]6294 The second argument is a pointer to a <tt>va_list</tt> element to copy
6295 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6296
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6297<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6298<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]6299 macro available in C. In a target-dependent way, it copies the
6300 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6301 element. This intrinsic is necessary because
6302 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6303 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6304
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6305</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6306
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6307</div>
6308
Bill Wendling0246bb72011-07-31 06:45:03 +0000[diff] [blame]6309</div>
6310
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6311<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6312<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6313 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6314</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6315
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6316<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6317
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6318<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]6319Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6320intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6321roots on the stack</a>, as well as garbage collector implementations that
6322require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6323barriers. Front-ends for type-safe garbage collected languages should generate
6324these intrinsics to make use of the LLVM garbage collectors. For more details,
6325see <a href="GarbageCollection.html">Accurate Garbage Collection with
6326LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6327
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6328<p>The garbage collection intrinsics only operate on objects in the generic
6329 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6330
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6331<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6332<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6333 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6334</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6335
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6336<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6337
6338<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6339<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6340 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6341</pre>
6342
6343<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]6344<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]6345 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6346
6347<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6348<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6349 root pointer. The second pointer (which must be either a constant or a
6350 global value address) contains the meta-data to be associated with the
6351 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6352
6353<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]6354<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]6355 location. At compile-time, the code generator generates information to allow
6356 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6357 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6358 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6359
6360</div>
6361
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6362<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6363<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6364 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6365</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6366
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6367<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6368
6369<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6370<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6371 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6372</pre>
6373
6374<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6375<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6376 locations, allowing garbage collector implementations that require read
6377 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6378
6379<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6380<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]6381 allocated from the garbage collector. The first object is a pointer to the
6382 start of the referenced object, if needed by the language runtime (otherwise
6383 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6384
6385<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6386<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6387 instruction, but may be replaced with substantially more complex code by the
6388 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6389 may only be used in a function which <a href="#gc">specifies a GC
6390 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6391
6392</div>
6393
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6394<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6395<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6396 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6397</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6398
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6399<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6400
6401<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6402<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6403 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6404</pre>
6405
6406<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6407<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6408 locations, allowing garbage collector implementations that require write
6409 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6410
6411<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6412<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]6413 object to store it to, and the third is the address of the field of Obj to
6414 store to. If the runtime does not require a pointer to the object, Obj may
6415 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6416
6417<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6418<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6419 instruction, but may be replaced with substantially more complex code by the
6420 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6421 may only be used in a function which <a href="#gc">specifies a GC
6422 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6423
6424</div>
6425
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6426</div>
6427
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6428<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6429<h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6430 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6431</h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6432
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6433<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6434
6435<p>These intrinsics are provided by LLVM to expose special features that may
6436 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6437
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6438<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6439<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6440 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6441</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6442
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6443<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6444
6445<h5>Syntax:</h5>
6446<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6447 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6448</pre>
6449
6450<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6451<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6452 target-specific value indicating the return address of the current function
6453 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6454
6455<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6456<p>The argument to this intrinsic indicates which function to return the address
6457 for. Zero indicates the calling function, one indicates its caller, etc.
6458 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6459
6460<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6461<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6462 indicating the return address of the specified call frame, or zero if it
6463 cannot be identified. The value returned by this intrinsic is likely to be
6464 incorrect or 0 for arguments other than zero, so it should only be used for
6465 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6466
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6467<p>Note that calling this intrinsic does not prevent function inlining or other
6468 aggressive transformations, so the value returned may not be that of the
6469 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6470
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6471</div>
6472
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6473<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6474<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6475 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6476</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6477
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6478<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6479
6480<h5>Syntax:</h5>
6481<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6482 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6483</pre>
6484
6485<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6486<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6487 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6488
6489<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6490<p>The argument to this intrinsic indicates which function to return the frame
6491 pointer for. Zero indicates the calling function, one indicates its caller,
6492 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6493
6494<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6495<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6496 indicating the frame address of the specified call frame, or zero if it
6497 cannot be identified. The value returned by this intrinsic is likely to be
6498 incorrect or 0 for arguments other than zero, so it should only be used for
6499 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6500
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6501<p>Note that calling this intrinsic does not prevent function inlining or other
6502 aggressive transformations, so the value returned may not be that of the
6503 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6504
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6505</div>
6506
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6507<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6508<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6509 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6510</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6511
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6512<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6513
6514<h5>Syntax:</h5>
6515<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6516 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6517</pre>
6518
6519<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6520<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6521 of the function stack, for use
6522 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6523 useful for implementing language features like scoped automatic variable
6524 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6525
6526<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6527<p>This intrinsic returns a opaque pointer value that can be passed
6528 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6529 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6530 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6531 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6532 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6533 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6534
6535</div>
6536
6537<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6538<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6539 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6540</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6541
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6542<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6543
6544<h5>Syntax:</h5>
6545<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6546 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6547</pre>
6548
6549<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6550<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6551 the function stack to the state it was in when the
6552 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6553 executed. This is useful for implementing language features like scoped
6554 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6555
6556<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6557<p>See the description
6558 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6559
6560</div>
6561
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6562<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6563<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6564 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6565</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6566
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6567<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6568
6569<h5>Syntax:</h5>
6570<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6571 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;, i32 &lt;cache type&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6572</pre>
6573
6574<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6575<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6576 insert a prefetch instruction if supported; otherwise, it is a noop.
6577 Prefetches have no effect on the behavior of the program but can change its
6578 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6579
6580<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6581<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6582 specifier determining if the fetch should be for a read (0) or write (1),
6583 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6584 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6585 specifies whether the prefetch is performed on the data (1) or instruction (0)
6586 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6587 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6588
6589<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6590<p>This intrinsic does not modify the behavior of the program. In particular,
6591 prefetches cannot trap and do not produce a value. On targets that support
6592 this intrinsic, the prefetch can provide hints to the processor cache for
6593 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6594
6595</div>
6596
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6597<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6598<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6599 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6600</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6601
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6602<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6603
6604<h5>Syntax:</h5>
6605<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6606 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6607</pre>
6608
6609<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6610<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6611 Counter (PC) in a region of code to simulators and other tools. The method
6612 is target specific, but it is expected that the marker will use exported
6613 symbols to transmit the PC of the marker. The marker makes no guarantees
6614 that it will remain with any specific instruction after optimizations. It is
6615 possible that the presence of a marker will inhibit optimizations. The
6616 intended use is to be inserted after optimizations to allow correlations of
6617 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6618
6619<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6620<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6621
6622<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6623<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6624 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6625
6626</div>
6627
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6628<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6629<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6630 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6631</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6632
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6633<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6634
6635<h5>Syntax:</h5>
6636<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]6637 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6638</pre>
6639
6640<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6641<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6642 counter register (or similar low latency, high accuracy clocks) on those
6643 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6644 should map to RPCC. As the backing counters overflow quickly (on the order
6645 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6646
6647<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6648<p>When directly supported, reading the cycle counter should not modify any
6649 memory. Implementations are allowed to either return a application specific
6650 value or a system wide value. On backends without support, this is lowered
6651 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6652
6653</div>
6654
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6655</div>
6656
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6657<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6658<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6659 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6660</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6661
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6662<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6663
6664<p>LLVM provides intrinsics for a few important standard C library functions.
6665 These intrinsics allow source-language front-ends to pass information about
6666 the alignment of the pointer arguments to the code generator, providing
6667 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6668
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6669<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6670<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6671 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6672</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6673
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6674<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6675
6676<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6677<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]6678 integer bit width and for different address spaces. Not all targets support
6679 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6680
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6681<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6682 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]6683 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6684 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]6685 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6686</pre>
6687
6688<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6689<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6690 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6691
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6692<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6693 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6694 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6695
6696<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6697
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6698<p>The first argument is a pointer to the destination, the second is a pointer
6699 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6700 number of bytes to copy, the fourth argument is the alignment of the
6701 source and destination locations, and the fifth is a boolean indicating a
6702 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6703
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6704<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]6705 then the caller guarantees that both the source and destination pointers are
6706 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6707
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6708<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6709 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6710 The detailed access behavior is not very cleanly specified and it is unwise
6711 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6712
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6713<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6714
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6715<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6716 source location to the destination location, which are not allowed to
6717 overlap. It copies "len" bytes of memory over. If the argument is known to
6718 be aligned to some boundary, this can be specified as the fourth argument,
6719 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6720
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6721</div>
6722
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6723<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6724<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6725 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6726</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6727
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6728<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6729
6730<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6731<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]6732 width and for different address space. Not all targets support all bit
6733 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6734
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6735<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6736 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]6737 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6738 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]6739 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6740</pre>
6741
6742<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6743<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6744 source location to the destination location. It is similar to the
6745 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6746 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6747
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6748<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6749 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6750 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6751
6752<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6753
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6754<p>The first argument is a pointer to the destination, the second is a pointer
6755 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6756 number of bytes to copy, the fourth argument is the alignment of the
6757 source and destination locations, and the fifth is a boolean indicating a
6758 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6759
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6760<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]6761 then the caller guarantees that the source and destination pointers are
6762 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6763
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6764<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6765 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6766 The detailed access behavior is not very cleanly specified and it is unwise
6767 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6768
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6769<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6770
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6771<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6772 source location to the destination location, which may overlap. It copies
6773 "len" bytes of memory over. If the argument is known to be aligned to some
6774 boundary, this can be specified as the fourth argument, otherwise it should
6775 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6776
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6777</div>
6778
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6779<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6780<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6781 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6782</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6783
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6784<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6785
6786<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6787<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]6788 width and for different address spaces. However, not all targets support all
6789 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6790
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6791<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6792 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6793 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6794 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6795 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6796</pre>
6797
6798<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6799<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6800 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6801
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6802<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6803 intrinsic does not return a value and takes extra alignment/volatile
6804 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6805
6806<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6807<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]6808 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6809 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6810 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6811
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6812<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]6813 then the caller guarantees that the destination pointer is aligned to that
6814 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6815
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6816<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6817 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6818 The detailed access behavior is not very cleanly specified and it is unwise
6819 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6820
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6821<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6822<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6823 at the destination location. If the argument is known to be aligned to some
6824 boundary, this can be specified as the fourth argument, otherwise it should
6825 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6826
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6827</div>
6828
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]6829<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6830<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6831 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6832</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6833
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6834<div>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6835
6836<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6837<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6838 floating point or vector of floating point type. Not all targets support all
6839 types however.</p>
6840
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6841<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]6842 declare float @llvm.sqrt.f32(float %Val)
6843 declare double @llvm.sqrt.f64(double %Val)
6844 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6845 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6846 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6847</pre>
6848
6849<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6850<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6851 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6852 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6853 behavior for negative numbers other than -0.0 (which allows for better
6854 optimization, because there is no need to worry about errno being
6855 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]6856
6857<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6858<p>The argument and return value are floating point numbers of the same
6859 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6860
6861<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6862<p>This function returns the sqrt of the specified operand if it is a
6863 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6864
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6865</div>
6866
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6867<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6868<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6869 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6870</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6871
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6872<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6873
6874<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6875<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6876 floating point or vector of floating point type. Not all targets support all
6877 types however.</p>
6878
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6879<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]6880 declare float @llvm.powi.f32(float %Val, i32 %power)
6881 declare double @llvm.powi.f64(double %Val, i32 %power)
6882 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6883 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6884 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6885</pre>
6886
6887<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6888<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6889 specified (positive or negative) power. The order of evaluation of
6890 multiplications is not defined. When a vector of floating point type is
6891 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6892
6893<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6894<p>The second argument is an integer power, and the first is a value to raise to
6895 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6896
6897<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6898<p>This function returns the first value raised to the second power with an
6899 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6900
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6901</div>
6902
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6903<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6904<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6905 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6906</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6907
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6908<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6909
6910<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6911<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6912 floating point or vector of floating point type. Not all targets support all
6913 types however.</p>
6914
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6915<pre>
6916 declare float @llvm.sin.f32(float %Val)
6917 declare double @llvm.sin.f64(double %Val)
6918 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6919 declare fp128 @llvm.sin.f128(fp128 %Val)
6920 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6921</pre>
6922
6923<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6924<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6925
6926<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6927<p>The argument and return value are floating point numbers of the same
6928 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6929
6930<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6931<p>This function returns the sine of the specified operand, returning the same
6932 values as the libm <tt>sin</tt> functions would, and handles error conditions
6933 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6934
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6935</div>
6936
6937<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6938<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6939 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6940</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6941
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6942<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6943
6944<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6945<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6946 floating point or vector of floating point type. Not all targets support all
6947 types however.</p>
6948
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6949<pre>
6950 declare float @llvm.cos.f32(float %Val)
6951 declare double @llvm.cos.f64(double %Val)
6952 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6953 declare fp128 @llvm.cos.f128(fp128 %Val)
6954 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6955</pre>
6956
6957<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6958<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6959
6960<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6961<p>The argument and return value are floating point numbers of the same
6962 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6963
6964<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6965<p>This function returns the cosine of the specified operand, returning the same
6966 values as the libm <tt>cos</tt> functions would, and handles error conditions
6967 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6968
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6969</div>
6970
6971<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6972<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6973 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6974</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6975
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6976<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6977
6978<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6979<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6980 floating point or vector of floating point type. Not all targets support all
6981 types however.</p>
6982
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6983<pre>
6984 declare float @llvm.pow.f32(float %Val, float %Power)
6985 declare double @llvm.pow.f64(double %Val, double %Power)
6986 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6987 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6988 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6989</pre>
6990
6991<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6992<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6993 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6994
6995<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6996<p>The second argument is a floating point power, and the first is a value to
6997 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6998
6999<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7000<p>This function returns the first value raised to the second power, returning
7001 the same values as the libm <tt>pow</tt> functions would, and handles error
7002 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7003
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7004</div>
7005
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7006</div>
7007
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7008<!-- _______________________________________________________________________ -->
7009<h4>
7010 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
7011</h4>
7012
7013<div>
7014
7015<h5>Syntax:</h5>
7016<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
7017 floating point or vector of floating point type. Not all targets support all
7018 types however.</p>
7019
7020<pre>
7021 declare float @llvm.exp.f32(float %Val)
7022 declare double @llvm.exp.f64(double %Val)
7023 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7024 declare fp128 @llvm.exp.f128(fp128 %Val)
7025 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7026</pre>
7027
7028<h5>Overview:</h5>
7029<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
7030
7031<h5>Arguments:</h5>
7032<p>The argument and return value are floating point numbers of the same
7033 type.</p>
7034
7035<h5>Semantics:</h5>
7036<p>This function returns the same values as the libm <tt>exp</tt> functions
7037 would, and handles error conditions in the same way.</p>
7038
7039</div>
7040
7041<!-- _______________________________________________________________________ -->
7042<h4>
7043 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
7044</h4>
7045
7046<div>
7047
7048<h5>Syntax:</h5>
7049<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
7050 floating point or vector of floating point type. Not all targets support all
7051 types however.</p>
7052
7053<pre>
7054 declare float @llvm.log.f32(float %Val)
7055 declare double @llvm.log.f64(double %Val)
7056 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7057 declare fp128 @llvm.log.f128(fp128 %Val)
7058 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7059</pre>
7060
7061<h5>Overview:</h5>
7062<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
7063
7064<h5>Arguments:</h5>
7065<p>The argument and return value are floating point numbers of the same
7066 type.</p>
7067
7068<h5>Semantics:</h5>
7069<p>This function returns the same values as the libm <tt>log</tt> functions
7070 would, and handles error conditions in the same way.</p>
7071
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7072<h4>
7073 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
7074</h4>
7075
7076<div>
7077
7078<h5>Syntax:</h5>
7079<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
7080 floating point or vector of floating point type. Not all targets support all
7081 types however.</p>
7082
7083<pre>
7084 declare float @llvm.fma.f32(float %a, float %b, float %c)
7085 declare double @llvm.fma.f64(double %a, double %b, double %c)
7086 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7087 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7088 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7089</pre>
7090
7091<h5>Overview:</h5>
Cameron Zwarichabc43e62011-07-08 22:13:55 +0000[diff] [blame]7092<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7093 operation.</p>
7094
7095<h5>Arguments:</h5>
7096<p>The argument and return value are floating point numbers of the same
7097 type.</p>
7098
7099<h5>Semantics:</h5>
7100<p>This function returns the same values as the libm <tt>fma</tt> functions
7101 would.</p>
7102
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7103</div>
7104
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7105<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7106<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7107 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7108</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7109
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7110<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7111
7112<p>LLVM provides intrinsics for a few important bit manipulation operations.
7113 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7114
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7115<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7116<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7117 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7118</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7119
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7120<div>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7121
7122<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7123<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7124 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
7125
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7126<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7127 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
7128 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
7129 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7130</pre>
7131
7132<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7133<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
7134 values with an even number of bytes (positive multiple of 16 bits). These
7135 are useful for performing operations on data that is not in the target's
7136 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7137
7138<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7139<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
7140 and low byte of the input i16 swapped. Similarly,
7141 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
7142 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7143 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7144 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7145 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7146 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7147
7148</div>
7149
7150<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7151<h4>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]7152 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7153</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7154
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7155<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7156
7157<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7158<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7159 width, or on any vector with integer elements. Not all targets support all
7160 bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7161
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7162<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7163 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7164 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]7165 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7166 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
7167 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7168 declare &lt;2 x i32&gt; @llvm.ctpop.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7169</pre>
7170
7171<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7172<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7173 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7174
7175<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7176<p>The only argument is the value to be counted. The argument may be of any
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7177 integer type, or a vector with integer elements.
7178 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7179
7180<h5>Semantics:</h5>
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7181<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7182 element of a vector.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7183
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7184</div>
7185
7186<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7187<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7188 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7189</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7190
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7191<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7192
7193<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7194<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7195 integer bit width, or any vector whose elements are integers. Not all
7196 targets support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7197
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7198<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7199 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
7200 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]7201 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7202 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
7203 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7204 declare &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src;gt)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7205</pre>
7206
7207<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7208<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7209 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7210
7211<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7212<p>The only argument is the value to be counted. The argument may be of any
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7213 integer type, or any vector type with integer element type.
7214 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7215
7216<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7217<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7218 zeros in a variable, or within each element of the vector if the operation
7219 is of vector type. If the src == 0 then the result is the size in bits of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7220 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7221
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7222</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7223
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7224<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7225<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7226 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7227</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7228
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7229<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7230
7231<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7232<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7233 integer bit width, or any vector of integer elements. Not all targets
7234 support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7235
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7236<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7237 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
7238 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]7239 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7240 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
7241 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7242 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7243</pre>
7244
7245<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7246<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7247 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7248
7249<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7250<p>The only argument is the value to be counted. The argument may be of any
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7251 integer type, or a vectory with integer element type.. The return type
7252 must match the argument type.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7253
7254<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7255<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7256 zeros in a variable, or within each element of a vector.
7257 If the src == 0 then the result is the size in bits of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7258 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7259
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7260</div>
7261
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7262</div>
7263
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7264<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7265<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7266 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7267</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7268
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7269<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7270
7271<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7272
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7273<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7274<h4>
7275 <a name="int_sadd_overflow">
7276 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7277 </a>
7278</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7279
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7280<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7281
7282<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7283<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]7284 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7285
7286<pre>
7287 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7288 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7289 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7290</pre>
7291
7292<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7293<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7294 a signed addition of the two arguments, and indicate whether an overflow
7295 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7296
7297<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7298<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]7299 be of integer types of any bit width, but they must have the same bit
7300 width. The second element of the result structure must be of
7301 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7302 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7303
7304<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7305<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7306 a signed addition of the two variables. They return a structure &mdash; the
7307 first element of which is the signed summation, and the second element of
7308 which is a bit specifying if the signed summation resulted in an
7309 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7310
7311<h5>Examples:</h5>
7312<pre>
7313 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7314 %sum = extractvalue {i32, i1} %res, 0
7315 %obit = extractvalue {i32, i1} %res, 1
7316 br i1 %obit, label %overflow, label %normal
7317</pre>
7318
7319</div>
7320
7321<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7322<h4>
7323 <a name="int_uadd_overflow">
7324 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7325 </a>
7326</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7327
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7328<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7329
7330<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7331<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]7332 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7333
7334<pre>
7335 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7336 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7337 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7338</pre>
7339
7340<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7341<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7342 an unsigned addition of the two arguments, and indicate whether a carry
7343 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7344
7345<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7346<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]7347 be of integer types of any bit width, but they must have the same bit
7348 width. The second element of the result structure must be of
7349 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7350 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7351
7352<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7353<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7354 an unsigned addition of the two arguments. They return a structure &mdash;
7355 the first element of which is the sum, and the second element of which is a
7356 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7357
7358<h5>Examples:</h5>
7359<pre>
7360 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7361 %sum = extractvalue {i32, i1} %res, 0
7362 %obit = extractvalue {i32, i1} %res, 1
7363 br i1 %obit, label %carry, label %normal
7364</pre>
7365
7366</div>
7367
7368<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7369<h4>
7370 <a name="int_ssub_overflow">
7371 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7372 </a>
7373</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7374
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7375<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7376
7377<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7378<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]7379 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7380
7381<pre>
7382 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7383 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7384 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7385</pre>
7386
7387<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7388<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7389 a signed subtraction of the two arguments, and indicate whether an overflow
7390 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7391
7392<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7393<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]7394 be of integer types of any bit width, but they must have the same bit
7395 width. The second element of the result structure must be of
7396 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7397 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7398
7399<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7400<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7401 a signed subtraction of the two arguments. They return a structure &mdash;
7402 the first element of which is the subtraction, and the second element of
7403 which is a bit specifying if the signed subtraction resulted in an
7404 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7405
7406<h5>Examples:</h5>
7407<pre>
7408 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7409 %sum = extractvalue {i32, i1} %res, 0
7410 %obit = extractvalue {i32, i1} %res, 1
7411 br i1 %obit, label %overflow, label %normal
7412</pre>
7413
7414</div>
7415
7416<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7417<h4>
7418 <a name="int_usub_overflow">
7419 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7420 </a>
7421</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7422
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7423<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7424
7425<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7426<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]7427 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7428
7429<pre>
7430 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7431 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7432 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7433</pre>
7434
7435<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7436<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7437 an unsigned subtraction of the two arguments, and indicate whether an
7438 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7439
7440<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7441<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]7442 be of integer types of any bit width, but they must have the same bit
7443 width. The second element of the result structure must be of
7444 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7445 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7446
7447<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7448<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7449 an unsigned subtraction of the two arguments. They return a structure &mdash;
7450 the first element of which is the subtraction, and the second element of
7451 which is a bit specifying if the unsigned subtraction resulted in an
7452 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7453
7454<h5>Examples:</h5>
7455<pre>
7456 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7457 %sum = extractvalue {i32, i1} %res, 0
7458 %obit = extractvalue {i32, i1} %res, 1
7459 br i1 %obit, label %overflow, label %normal
7460</pre>
7461
7462</div>
7463
7464<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7465<h4>
7466 <a name="int_smul_overflow">
7467 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7468 </a>
7469</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7470
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7471<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7472
7473<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7474<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]7475 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7476
7477<pre>
7478 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7479 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7480 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7481</pre>
7482
7483<h5>Overview:</h5>
7484
7485<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7486 a signed multiplication of the two arguments, and indicate whether an
7487 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7488
7489<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7490<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]7491 be of integer types of any bit width, but they must have the same bit
7492 width. The second element of the result structure must be of
7493 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7494 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7495
7496<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7497<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7498 a signed multiplication of the two arguments. They return a structure &mdash;
7499 the first element of which is the multiplication, and the second element of
7500 which is a bit specifying if the signed multiplication resulted in an
7501 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7502
7503<h5>Examples:</h5>
7504<pre>
7505 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7506 %sum = extractvalue {i32, i1} %res, 0
7507 %obit = extractvalue {i32, i1} %res, 1
7508 br i1 %obit, label %overflow, label %normal
7509</pre>
7510
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]7511</div>
7512
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7513<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7514<h4>
7515 <a name="int_umul_overflow">
7516 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7517 </a>
7518</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7519
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7520<div>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7521
7522<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7523<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]7524 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7525
7526<pre>
7527 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7528 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7529 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7530</pre>
7531
7532<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7533<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7534 a unsigned multiplication of the two arguments, and indicate whether an
7535 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7536
7537<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7538<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]7539 be of integer types of any bit width, but they must have the same bit
7540 width. The second element of the result structure must be of
7541 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7542 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7543
7544<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7545<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7546 an unsigned multiplication of the two arguments. They return a structure
7547 &mdash; the first element of which is the multiplication, and the second
7548 element of which is a bit specifying if the unsigned multiplication resulted
7549 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7550
7551<h5>Examples:</h5>
7552<pre>
7553 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7554 %sum = extractvalue {i32, i1} %res, 0
7555 %obit = extractvalue {i32, i1} %res, 1
7556 br i1 %obit, label %overflow, label %normal
7557</pre>
7558
7559</div>
7560
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7561</div>
7562
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7563<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7564<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7565 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7566</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7567
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7568<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7569
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7570<p>Half precision floating point is a storage-only format. This means that it is
7571 a dense encoding (in memory) but does not support computation in the
7572 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7573
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7574<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7575 value as an i16, then convert it to float with <a
7576 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7577 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7578 double etc). To store the value back to memory, it is first converted to
7579 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7580 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7581 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7582
7583<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7584<h4>
7585 <a name="int_convert_to_fp16">
7586 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7587 </a>
7588</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7589
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7590<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7591
7592<h5>Syntax:</h5>
7593<pre>
7594 declare i16 @llvm.convert.to.fp16(f32 %a)
7595</pre>
7596
7597<h5>Overview:</h5>
7598<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7599 a conversion from single precision floating point format to half precision
7600 floating point format.</p>
7601
7602<h5>Arguments:</h5>
7603<p>The intrinsic function contains single argument - the value to be
7604 converted.</p>
7605
7606<h5>Semantics:</h5>
7607<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7608 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7609 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7610 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7611
7612<h5>Examples:</h5>
7613<pre>
7614 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7615 store i16 %res, i16* @x, align 2
7616</pre>
7617
7618</div>
7619
7620<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7621<h4>
7622 <a name="int_convert_from_fp16">
7623 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7624 </a>
7625</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7626
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7627<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7628
7629<h5>Syntax:</h5>
7630<pre>
7631 declare f32 @llvm.convert.from.fp16(i16 %a)
7632</pre>
7633
7634<h5>Overview:</h5>
7635<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7636 a conversion from half precision floating point format to single precision
7637 floating point format.</p>
7638
7639<h5>Arguments:</h5>
7640<p>The intrinsic function contains single argument - the value to be
7641 converted.</p>
7642
7643<h5>Semantics:</h5>
7644<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7645 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7646 precision floating point format. The input half-float value is represented by
7647 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7648
7649<h5>Examples:</h5>
7650<pre>
7651 %a = load i16* @x, align 2
7652 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7653</pre>
7654
7655</div>
7656
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7657</div>
7658
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7659<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7660<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7661 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7662</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7663
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7664<div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7665
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7666<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7667 prefix), are described in
7668 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7669 Level Debugging</a> document.</p>
7670
7671</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7672
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7673<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7674<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7675 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7676</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7677
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7678<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7679
7680<p>The LLVM exception handling intrinsics (which all start with
7681 <tt>llvm.eh.</tt> prefix), are described in
7682 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7683 Handling</a> document.</p>
7684
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7685</div>
7686
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7687<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7688<h3>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7689 <a name="int_trampoline">Trampoline Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7690</h3>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7691
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7692<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7693
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7694<p>These intrinsics make it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]7695 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7696 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7697 function pointer lacking the nest parameter - the caller does not need to
7698 provide a value for it. Instead, the value to use is stored in advance in a
7699 "trampoline", a block of memory usually allocated on the stack, which also
7700 contains code to splice the nest value into the argument list. This is used
7701 to implement the GCC nested function address extension.</p>
7702
7703<p>For example, if the function is
7704 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7705 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7706 follows:</p>
7707
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]7708<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]7709 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7710 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7711 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
7712 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]7713 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7714</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7715
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7716<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7717 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7718
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7719<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7720<h4>
7721 <a name="int_it">
7722 '<tt>llvm.init.trampoline</tt>' Intrinsic
7723 </a>
7724</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7725
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7726<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7727
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7728<h5>Syntax:</h5>
7729<pre>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7730 declare void @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]7731</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7732
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7733<h5>Overview:</h5>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7734<p>This fills the memory pointed to by <tt>tramp</tt> with executable code,
7735 turning it into a trampoline.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7736
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7737<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7738<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7739 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7740 sufficiently aligned block of memory; this memory is written to by the
7741 intrinsic. Note that the size and the alignment are target-specific - LLVM
7742 currently provides no portable way of determining them, so a front-end that
7743 generates this intrinsic needs to have some target-specific knowledge.
7744 The <tt>func</tt> argument must hold a function bitcast to
7745 an <tt>i8*</tt>.</p>
7746
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7747<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7748<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7749 dependent code, turning it into a function. Then <tt>tramp</tt> needs to be
7750 passed to <a href="#int_at">llvm.adjust.trampoline</a> to get a pointer
7751 which can be <a href="#int_trampoline">bitcast (to a new function) and
7752 called</a>. The new function's signature is the same as that of
7753 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
7754 removed. At most one such <tt>nest</tt> argument is allowed, and it must be of
7755 pointer type. Calling the new function is equivalent to calling <tt>func</tt>
7756 with the same argument list, but with <tt>nval</tt> used for the missing
7757 <tt>nest</tt> argument. If, after calling <tt>llvm.init.trampoline</tt>, the
7758 memory pointed to by <tt>tramp</tt> is modified, then the effect of any later call
7759 to the returned function pointer is undefined.</p>
7760</div>
7761
7762<!-- _______________________________________________________________________ -->
7763<h4>
7764 <a name="int_at">
7765 '<tt>llvm.adjust.trampoline</tt>' Intrinsic
7766 </a>
7767</h4>
7768
7769<div>
7770
7771<h5>Syntax:</h5>
7772<pre>
7773 declare i8* @llvm.adjust.trampoline(i8* &lt;tramp&gt;)
7774</pre>
7775
7776<h5>Overview:</h5>
7777<p>This performs any required machine-specific adjustment to the address of a
7778 trampoline (passed as <tt>tramp</tt>).</p>
7779
7780<h5>Arguments:</h5>
7781<p><tt>tramp</tt> must point to a block of memory which already has trampoline code
7782 filled in by a previous call to <a href="#int_it"><tt>llvm.init.trampoline</tt>
7783 </a>.</p>
7784
7785<h5>Semantics:</h5>
7786<p>On some architectures the address of the code to be executed needs to be
7787 different to the address where the trampoline is actually stored. This
7788 intrinsic returns the executable address corresponding to <tt>tramp</tt>
7789 after performing the required machine specific adjustments.
7790 The pointer returned can then be <a href="#int_trampoline"> bitcast and
7791 executed</a>.
7792</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7793
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7794</div>
7795
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7796</div>
7797
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7798<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7799<h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7800 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7801</h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7802
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7803<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7804
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7805<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7806 hardware constructs for atomic operations and memory synchronization. This
7807 provides an interface to the hardware, not an interface to the programmer. It
7808 is aimed at a low enough level to allow any programming models or APIs
7809 (Application Programming Interfaces) which need atomic behaviors to map
7810 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7811 hardware provides a "universal IR" for source languages, it also provides a
7812 starting point for developing a "universal" atomic operation and
7813 synchronization IR.</p>
7814
7815<p>These do <em>not</em> form an API such as high-level threading libraries,
7816 software transaction memory systems, atomic primitives, and intrinsic
7817 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7818 application libraries. The hardware interface provided by LLVM should allow
7819 a clean implementation of all of these APIs and parallel programming models.
7820 No one model or paradigm should be selected above others unless the hardware
7821 itself ubiquitously does so.</p>
7822
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7823<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7824<h4>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7825 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7826</h4>
7827
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7828<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7829<h5>Syntax:</h5>
7830<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7831 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]7832</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7833
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7834<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7835<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7836 specific pairs of memory access types.</p>
7837
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7838<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7839<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7840 The first four arguments enables a specific barrier as listed below. The
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7841 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7842 memory.</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7843
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7844<ul>
7845 <li><tt>ll</tt>: load-load barrier</li>
7846 <li><tt>ls</tt>: load-store barrier</li>
7847 <li><tt>sl</tt>: store-load barrier</li>
7848 <li><tt>ss</tt>: store-store barrier</li>
7849 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7850</ul>
7851
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7852<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7853<p>This intrinsic causes the system to enforce some ordering constraints upon
7854 the loads and stores of the program. This barrier does not
7855 indicate <em>when</em> any events will occur, it only enforces
7856 an <em>order</em> in which they occur. For any of the specified pairs of load
7857 and store operations (f.ex. load-load, or store-load), all of the first
7858 operations preceding the barrier will complete before any of the second
7859 operations succeeding the barrier begin. Specifically the semantics for each
7860 pairing is as follows:</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7861
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7862<ul>
7863 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7864 after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7865 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7866 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7867 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7868 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7869 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7870 load after the barrier begins.</li>
7871</ul>
7872
7873<p>These semantics are applied with a logical "and" behavior when more than one
7874 is enabled in a single memory barrier intrinsic.</p>
7875
7876<p>Backends may implement stronger barriers than those requested when they do
7877 not support as fine grained a barrier as requested. Some architectures do
7878 not need all types of barriers and on such architectures, these become
7879 noops.</p>
7880
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7881<h5>Example:</h5>
7882<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7883%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7884%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7885 store i32 4, %ptr
7886
7887%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Evan Cheng0b0669a2011-06-29 17:14:00 +0000[diff] [blame]7888 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false, i1 true)
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7889 <i>; guarantee the above finishes</i>
7890 store i32 8, %ptr <i>; before this begins</i>
7891</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7892
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7893</div>
7894
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7895<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7896<h4>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7897 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7898</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7899
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7900<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7901
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7902<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7903<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7904 any integer bit width and for different address spaces. Not all targets
7905 support all bit widths however.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7906
7907<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7908 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7909 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7910 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7911 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]7912</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7913
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7914<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7915<p>This loads a value in memory and compares it to a given value. If they are
7916 equal, it stores a new value into the memory.</p>
7917
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7918<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7919<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7920 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7921 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7922 this integer type. While any bit width integer may be used, targets may only
7923 lower representations they support in hardware.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7924
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7925<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7926<p>This entire intrinsic must be executed atomically. It first loads the value
7927 in memory pointed to by <tt>ptr</tt> and compares it with the
7928 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7929 memory. The loaded value is yielded in all cases. This provides the
7930 equivalent of an atomic compare-and-swap operation within the SSA
7931 framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7932
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7933<h5>Examples:</h5>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7934<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7935%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7936%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7937 store i32 4, %ptr
7938
7939%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7940%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7941 <i>; yields {i32}:result1 = 4</i>
7942%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7943%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7944
7945%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7946%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7947 <i>; yields {i32}:result2 = 8</i>
7948%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7949
7950%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7951</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7952
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7953</div>
7954
7955<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7956<h4>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7957 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7958</h4>
7959
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7960<div>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7961<h5>Syntax:</h5>
7962
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7963<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7964 integer bit width. Not all targets support all bit widths however.</p>
7965
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7966<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7967 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7968 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7969 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7970 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]7971</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7972
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7973<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7974<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7975 the value from memory. It then stores the value in <tt>val</tt> in the memory
7976 at <tt>ptr</tt>.</p>
7977
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7978<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7979<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7980 the <tt>val</tt> argument and the result must be integers of the same bit
7981 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7982 integer type. The targets may only lower integer representations they
7983 support.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7984
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7985<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7986<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7987 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7988 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7989
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7990<h5>Examples:</h5>
7991<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7992%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7993%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7994 store i32 4, %ptr
7995
7996%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7997%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7998 <i>; yields {i32}:result1 = 4</i>
7999%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
8000%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
8001
8002%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8003%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8004 <i>; yields {i32}:result2 = 8</i>
8005
8006%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
8007%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
8008</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8009
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8010</div>
8011
8012<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8013<h4>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8014 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8015</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8016
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8017<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8018
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8019<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8020<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
8021 any integer bit width. Not all targets support all bit widths however.</p>
8022
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8023<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8024 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8025 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8026 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8027 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]8028</pre>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8029
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8030<h5>Overview:</h5>
8031<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
8032 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
8033
8034<h5>Arguments:</h5>
8035<p>The intrinsic takes two arguments, the first a pointer to an integer value
8036 and the second an integer value. The result is also an integer value. These
8037 integer types can have any bit width, but they must all have the same bit
8038 width. The targets may only lower integer representations they support.</p>
8039
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8040<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8041<p>This intrinsic does a series of operations atomically. It first loads the
8042 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
8043 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]8044
8045<h5>Examples:</h5>
8046<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]8047%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8048%ptr = bitcast i8* %mallocP to i32*
8049 store i32 4, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8050%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8051 <i>; yields {i32}:result1 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8052%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8053 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8054%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8055 <i>; yields {i32}:result3 = 10</i>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8056%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8057</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8058
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]8059</div>
8060
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8061<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8062<h4>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8063 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8064</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8065
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8066<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8067
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8068<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8069<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
8070 any integer bit width and for different address spaces. Not all targets
8071 support all bit widths however.</p>
8072
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8073<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8074 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8075 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8076 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8077 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]8078</pre>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8079
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8080<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]8081<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8082 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
8083
8084<h5>Arguments:</h5>
8085<p>The intrinsic takes two arguments, the first a pointer to an integer value
8086 and the second an integer value. The result is also an integer value. These
8087 integer types can have any bit width, but they must all have the same bit
8088 width. The targets may only lower integer representations they support.</p>
8089
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8090<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8091<p>This intrinsic does a series of operations atomically. It first loads the
8092 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
8093 result to <tt>ptr</tt>. It yields the original value stored
8094 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8095
8096<h5>Examples:</h5>
8097<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]8098%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8099%ptr = bitcast i8* %mallocP to i32*
8100 store i32 8, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8101%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8102 <i>; yields {i32}:result1 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8103%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8104 <i>; yields {i32}:result2 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8105%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8106 <i>; yields {i32}:result3 = 2</i>
8107%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
8108</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8109
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8110</div>
8111
8112<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8113<h4>
8114 <a name="int_atomic_load_and">
8115 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
8116 </a>
8117 <br>
8118 <a name="int_atomic_load_nand">
8119 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
8120 </a>
8121 <br>
8122 <a name="int_atomic_load_or">
8123 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
8124 </a>
8125 <br>
8126 <a name="int_atomic_load_xor">
8127 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
8128 </a>
8129</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8130
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8131<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8132
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8133<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8134<p>These are overloaded intrinsics. You can
8135 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
8136 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
8137 bit width and for different address spaces. Not all targets support all bit
8138 widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8139
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8140<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8141 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8142 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8143 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8144 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]8145</pre>
8146
8147<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8148 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8149 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8150 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8151 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]8152</pre>
8153
8154<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8155 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8156 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8157 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8158 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]8159</pre>
8160
8161<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8162 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8163 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8164 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8165 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]8166</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8167
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8168<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8169<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
8170 the value stored in memory at <tt>ptr</tt>. It yields the original value
8171 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8172
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8173<h5>Arguments:</h5>
8174<p>These intrinsics take two arguments, the first a pointer to an integer value
8175 and the second an integer value. The result is also an integer value. These
8176 integer types can have any bit width, but they must all have the same bit
8177 width. The targets may only lower integer representations they support.</p>
8178
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8179<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8180<p>These intrinsics does a series of operations atomically. They first load the
8181 value stored at <tt>ptr</tt>. They then do the bitwise
8182 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
8183 original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8184
8185<h5>Examples:</h5>
8186<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]8187%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8188%ptr = bitcast i8* %mallocP to i32*
8189 store i32 0x0F0F, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8190%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8191 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8192%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8193 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8194%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8195 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8196%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8197 <i>; yields {i32}:result3 = FF</i>
8198%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
8199</pre>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8200
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8201</div>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8202
8203<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8204<h4>
8205 <a name="int_atomic_load_max">
8206 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
8207 </a>
8208 <br>
8209 <a name="int_atomic_load_min">
8210 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
8211 </a>
8212 <br>
8213 <a name="int_atomic_load_umax">
8214 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
8215 </a>
8216 <br>
8217 <a name="int_atomic_load_umin">
8218 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
8219 </a>
8220</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8221
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8222<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8223
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8224<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8225<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
8226 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
8227 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
8228 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8229
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8230<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8231 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8232 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8233 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8234 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]8235</pre>
8236
8237<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8238 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8239 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8240 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8241 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]8242</pre>
8243
8244<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8245 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8246 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8247 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8248 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]8249</pre>
8250
8251<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8252 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
8253 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
8254 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
8255 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]8256</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8257
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8258<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]8259<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8260 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
8261 original value at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8262
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8263<h5>Arguments:</h5>
8264<p>These intrinsics take two arguments, the first a pointer to an integer value
8265 and the second an integer value. The result is also an integer value. These
8266 integer types can have any bit width, but they must all have the same bit
8267 width. The targets may only lower integer representations they support.</p>
8268
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8269<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8270<p>These intrinsics does a series of operations atomically. They first load the
8271 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
8272 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
8273 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8274
8275<h5>Examples:</h5>
8276<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]8277%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
8278%ptr = bitcast i8* %mallocP to i32*
8279 store i32 7, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8280%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8281 <i>; yields {i32}:result0 = 7</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8282%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8283 <i>; yields {i32}:result1 = -2</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8284%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8285 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8286%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8287 <i>; yields {i32}:result3 = 8</i>
8288%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
8289</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8290
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]8291</div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]8292
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8293</div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8294
8295<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8296<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8297 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8298</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8299
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8300<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8301
8302<p>This class of intrinsics exists to information about the lifetime of memory
8303 objects and ranges where variables are immutable.</p>
8304
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8305<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8306<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8307 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8308</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8309
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8310<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8311
8312<h5>Syntax:</h5>
8313<pre>
8314 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8315</pre>
8316
8317<h5>Overview:</h5>
8318<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
8319 object's lifetime.</p>
8320
8321<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8322<p>The first argument is a constant integer representing the size of the
8323 object, or -1 if it is variable sized. The second argument is a pointer to
8324 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8325
8326<h5>Semantics:</h5>
8327<p>This intrinsic indicates that before this point in the code, the value of the
8328 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]8329 never be used and has an undefined value. A load from the pointer that
8330 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8331 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
8332
8333</div>
8334
8335<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8336<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8337 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8338</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8339
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8340<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8341
8342<h5>Syntax:</h5>
8343<pre>
8344 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8345</pre>
8346
8347<h5>Overview:</h5>
8348<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
8349 object's lifetime.</p>
8350
8351<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8352<p>The first argument is a constant integer representing the size of the
8353 object, or -1 if it is variable sized. The second argument is a pointer to
8354 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8355
8356<h5>Semantics:</h5>
8357<p>This intrinsic indicates that after this point in the code, the value of the
8358 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8359 never be used and has an undefined value. Any stores into the memory object
8360 following this intrinsic may be removed as dead.
8361
8362</div>
8363
8364<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8365<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8366 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8367</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8368
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8369<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8370
8371<h5>Syntax:</h5>
8372<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +0000[diff] [blame]8373 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8374</pre>
8375
8376<h5>Overview:</h5>
8377<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8378 a memory object will not change.</p>
8379
8380<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8381<p>The first argument is a constant integer representing the size of the
8382 object, or -1 if it is variable sized. The second argument is a pointer to
8383 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8384
8385<h5>Semantics:</h5>
8386<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8387 the return value, the referenced memory location is constant and
8388 unchanging.</p>
8389
8390</div>
8391
8392<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8393<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8394 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8395</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8396
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8397<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8398
8399<h5>Syntax:</h5>
8400<pre>
8401 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8402</pre>
8403
8404<h5>Overview:</h5>
8405<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8406 a memory object are mutable.</p>
8407
8408<h5>Arguments:</h5>
8409<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8410 The second argument is a constant integer representing the size of the
8411 object, or -1 if it is variable sized and the third argument is a pointer
8412 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8413
8414<h5>Semantics:</h5>
8415<p>This intrinsic indicates that the memory is mutable again.</p>
8416
8417</div>
8418
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8419</div>
8420
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]8421<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8422<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8423 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8424</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8425
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8426<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8427
8428<p>This class of intrinsics is designed to be generic and has no specific
8429 purpose.</p>
8430
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8431<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8432<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8433 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8434</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8435
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8436<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8437
8438<h5>Syntax:</h5>
8439<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8440 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]8441</pre>
8442
8443<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8444<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8445
8446<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8447<p>The first argument is a pointer to a value, the second is a pointer to a
8448 global string, the third is a pointer to a global string which is the source
8449 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8450
8451<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8452<p>This intrinsic allows annotation of local variables with arbitrary strings.
8453 This can be useful for special purpose optimizations that want to look for
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8454 these annotations. These have no other defined use; they are ignored by code
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8455 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8456
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8457</div>
8458
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8459<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8460<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]8461 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8462</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8463
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8464<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8465
8466<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8467<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8468 any integer bit width.</p>
8469
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8470<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8471 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8472 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8473 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8474 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8475 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]8476</pre>
8477
8478<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8479<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8480
8481<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8482<p>The first argument is an integer value (result of some expression), the
8483 second is a pointer to a global string, the third is a pointer to a global
8484 string which is the source file name, and the last argument is the line
8485 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8486
8487<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8488<p>This intrinsic allows annotations to be put on arbitrary expressions with
8489 arbitrary strings. This can be useful for special purpose optimizations that
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8490 want to look for these annotations. These have no other defined use; they
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8491 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8492
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8493</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8494
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8495<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8496<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8497 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8498</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8499
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8500<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8501
8502<h5>Syntax:</h5>
8503<pre>
8504 declare void @llvm.trap()
8505</pre>
8506
8507<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8508<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8509
8510<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8511<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8512
8513<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8514<p>This intrinsics is lowered to the target dependent trap instruction. If the
8515 target does not have a trap instruction, this intrinsic will be lowered to
8516 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8517
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8518</div>
8519
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8520<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8521<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +0000[diff] [blame]8522 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8523</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8524
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8525<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8526
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8527<h5>Syntax:</h5>
8528<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8529 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8530</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8531
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8532<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8533<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8534 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8535 ensure that it is placed on the stack before local variables.</p>
8536
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8537<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8538<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8539 arguments. The first argument is the value loaded from the stack
8540 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8541 that has enough space to hold the value of the guard.</p>
8542
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8543<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8544<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8545 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8546 stack. This is to ensure that if a local variable on the stack is
8547 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8548 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8549 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8550 function.</p>
8551
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8552</div>
8553
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8554<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8555<h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8556 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8557</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8558
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8559<div>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8560
8561<h5>Syntax:</h5>
8562<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8563 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8564 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8565</pre>
8566
8567<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8568<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8569 the optimizers to determine at compile time whether a) an operation (like
8570 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8571 runtime check for overflow isn't necessary. An object in this context means
8572 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8573
8574<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8575<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8576 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8577 is a boolean 0 or 1. This argument determines whether you want the
8578 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]8579 1, variables are not allowed.</p>
8580
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8581<h5>Semantics:</h5>
8582<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8583 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8584 depending on the <tt>type</tt> argument, if the size cannot be determined at
8585 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8586
8587</div>
8588
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8589</div>
8590
8591</div>
8592
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8593<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8594<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8595<address>
8596 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8597 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8598 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8599 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]8600
8601 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]8602 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8603 Last modified: $Date$
8604</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]8605
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]8606</body>
8607</html>