Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 88a680837768d370d794960b677438d45fde148e / . / docs / LangRef.html
blob: 29ac260fd0ed7c506fae62fca36724919f4ca323 [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>
Bill Wendlingf7f06102011-10-11 06:41:28 +0000[diff] [blame]38 <li><a href="#linkage_external">'<tt>external</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]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 Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]95 <li><a href="#poisonvalues">Poison 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>
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a>
104 <ol>
105 <li><a href="#tbaa">'<tt>tbaa</tt>' Metadata</a></li>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]106 <li><a href="#fpaccuracy">'<tt>fpaccuracy</tt>' Metadata</a></li>
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]107 </ol>
108 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]109 </ol>
110 </li>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]111 <li><a href="#module_flags">Module Flags Metadata</a>
112 <ol>
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]113 <li><a href="#objc_gc_flags">Objective-C Garbage Collection Module Flags Metadata</a></li>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]114 </ol>
115 </li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]116 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
117 <ol>
118 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]119 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
120 Global Variable</a></li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]121 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
122 Global Variable</a></li>
123 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
124 Global Variable</a></li>
125 </ol>
126 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]127 <li><a href="#instref">Instruction Reference</a>
128 <ol>
129 <li><a href="#terminators">Terminator Instructions</a>
130 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]131 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
132 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]133 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]134 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]135 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]136 <li><a href="#i_resume">'<tt>resume</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]137 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]138 </ol>
139 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]140 <li><a href="#binaryops">Binary Operations</a>
141 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]142 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]143 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]144 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]145 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]146 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]147 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]148 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
149 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
150 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]151 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
152 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
153 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]154 </ol>
155 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]156 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
157 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]158 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
159 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
160 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]161 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]162 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]163 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]164 </ol>
165 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]166 <li><a href="#vectorops">Vector Operations</a>
167 <ol>
168 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
169 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
170 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]171 </ol>
172 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]173 <li><a href="#aggregateops">Aggregate Operations</a>
174 <ol>
175 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
176 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
177 </ol>
178 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]179 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]180 <ol>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]181 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
182 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
183 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
184 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
185 <li><a href="#i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a></li>
186 <li><a href="#i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]187 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]188 </ol>
189 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]190 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]191 <ol>
192 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
193 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
194 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
195 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
196 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]197 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
198 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
199 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
200 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]201 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
202 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]203 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]204 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]205 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]206 <li><a href="#otherops">Other Operations</a>
207 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]208 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
209 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]210 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]211 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]212 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]213 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]214 <li><a href="#i_landingpad">'<tt>landingpad</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]215 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]216 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]217 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]218 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]219 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]220 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]221 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
222 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]223 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
224 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
225 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]226 </ol>
227 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]228 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
229 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]230 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
231 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
232 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]233 </ol>
234 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]235 <li><a href="#int_codegen">Code Generator Intrinsics</a>
236 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]237 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
238 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
239 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
240 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
241 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
242 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000[diff] [blame]243 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]244 </ol>
245 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]246 <li><a href="#int_libc">Standard C Library Intrinsics</a>
247 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]248 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
249 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
250 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
251 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
252 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]253 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
254 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
255 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohman08b280b2011-05-27 00:36:31 +0000[diff] [blame]256 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
257 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]258 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]259 </ol>
260 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]261 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]262 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]263 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]264 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
265 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
266 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]267 </ol>
268 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]269 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
270 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]271 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
272 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
273 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
274 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
275 <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]276 <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]277 </ol>
278 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]279 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
280 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]281 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
282 <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]283 </ol>
284 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]285 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]286 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]287 <li><a href="#int_trampoline">Trampoline Intrinsics</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]288 <ol>
289 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]290 <li><a href="#int_at">'<tt>llvm.adjust.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]291 </ol>
292 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]293 <li><a href="#int_memorymarkers">Memory Use Markers</a>
294 <ol>
Jakub Staszak8e1b12a2011-12-04 20:44:25 +0000[diff] [blame]295 <li><a href="#int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a></li>
296 <li><a href="#int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a></li>
297 <li><a href="#int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a></li>
298 <li><a href="#int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a></li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]299 </ol>
300 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]301 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]302 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]303 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]304 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]305 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]306 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]307 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]308 '<tt>llvm.trap</tt>' Intrinsic</a></li>
309 <li><a href="#int_stackprotector">
310 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]311 <li><a href="#int_objectsize">
312 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]313 <li><a href="#int_expect">
314 '<tt>llvm.expect</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]315 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]316 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]317 </ol>
318 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]319</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]320
321<div class="doc_author">
322 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
323 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]324</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]325
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]326<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]327<h2><a name="abstract">Abstract</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]328<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]329
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]330<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]331
332<p>This document is a reference manual for the LLVM assembly language. LLVM is
333 a Static Single Assignment (SSA) based representation that provides type
334 safety, low-level operations, flexibility, and the capability of representing
335 'all' high-level languages cleanly. It is the common code representation
336 used throughout all phases of the LLVM compilation strategy.</p>
337
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]338</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]339
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]340<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]341<h2><a name="introduction">Introduction</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]342<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]343
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]344<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]345
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]346<p>The LLVM code representation is designed to be used in three different forms:
347 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
348 for fast loading by a Just-In-Time compiler), and as a human readable
349 assembly language representation. This allows LLVM to provide a powerful
350 intermediate representation for efficient compiler transformations and
351 analysis, while providing a natural means to debug and visualize the
352 transformations. The three different forms of LLVM are all equivalent. This
353 document describes the human readable representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]354
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]355<p>The LLVM representation aims to be light-weight and low-level while being
356 expressive, typed, and extensible at the same time. It aims to be a
357 "universal IR" of sorts, by being at a low enough level that high-level ideas
358 may be cleanly mapped to it (similar to how microprocessors are "universal
359 IR's", allowing many source languages to be mapped to them). By providing
360 type information, LLVM can be used as the target of optimizations: for
361 example, through pointer analysis, it can be proven that a C automatic
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]362 variable is never accessed outside of the current function, allowing it to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]363 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]364
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]365<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]366<h4>
367 <a name="wellformed">Well-Formedness</a>
368</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]369
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]370<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]371
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]372<p>It is important to note that this document describes 'well formed' LLVM
373 assembly language. There is a difference between what the parser accepts and
374 what is considered 'well formed'. For example, the following instruction is
375 syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]376
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]377<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]378%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]379</pre>
380
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]381<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
382 LLVM infrastructure provides a verification pass that may be used to verify
383 that an LLVM module is well formed. This pass is automatically run by the
384 parser after parsing input assembly and by the optimizer before it outputs
385 bitcode. The violations pointed out by the verifier pass indicate bugs in
386 transformation passes or input to the parser.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]387
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]388</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]389
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]390</div>
391
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame]392<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]393
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]394<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]395<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]396<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]397
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]398<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]399
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]400<p>LLVM identifiers come in two basic types: global and local. Global
401 identifiers (functions, global variables) begin with the <tt>'@'</tt>
402 character. Local identifiers (register names, types) begin with
403 the <tt>'%'</tt> character. Additionally, there are three different formats
404 for identifiers, for different purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]405
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]406<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]407 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]408 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
409 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
410 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
411 other characters in their names can be surrounded with quotes. Special
412 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
413 ASCII code for the character in hexadecimal. In this way, any character
414 can be used in a name value, even quotes themselves.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]415
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]416 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]417 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]418
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]419 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]420 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]421</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]422
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]423<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]424 don't need to worry about name clashes with reserved words, and the set of
425 reserved words may be expanded in the future without penalty. Additionally,
426 unnamed identifiers allow a compiler to quickly come up with a temporary
427 variable without having to avoid symbol table conflicts.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]428
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]429<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]430 languages. There are keywords for different opcodes
431 ('<tt><a href="#i_add">add</a></tt>',
432 '<tt><a href="#i_bitcast">bitcast</a></tt>',
433 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
434 ('<tt><a href="#t_void">void</a></tt>',
435 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
436 reserved words cannot conflict with variable names, because none of them
437 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]438
439<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]440 '<tt>%X</tt>' by 8:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]441
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]442<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]443
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]444<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]445%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]446</pre>
447
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]448<p>After strength reduction:</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_shl">shl</a> i32 %X, i8 3
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>And the hard way:</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">
Gabor Greifec58f752009-10-28 13:05:07 +0000[diff] [blame]457%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
458%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]459%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]460</pre>
461
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]462<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
463 lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]464
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]465<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]466 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]467 line.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]468
469 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]470 assigned to a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]471
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]472 <li>Unnamed temporaries are numbered sequentially</li>
473</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]474
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]475<p>It also shows a convention that we follow in this document. When
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]476 demonstrating instructions, we will follow an instruction with a comment that
477 defines the type and name of value produced. Comments are shown in italic
478 text.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]479
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]480</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]481
482<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]483<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]484<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]485<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]486<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]487<h3>
488 <a name="modulestructure">Module Structure</a>
489</h3>
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
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]493<p>LLVM programs are composed of "Module"s, each of which is a translation unit
494 of the input programs. Each module consists of functions, global variables,
495 and symbol table entries. Modules may be combined together with the LLVM
496 linker, which merges function (and global variable) definitions, resolves
497 forward declarations, and merges symbol table entries. Here is an example of
498 the "hello world" module:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]499
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]500<pre class="doc_code">
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]501<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckydb9cd762011-01-29 01:09:53 +0000[diff] [blame]502<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]503
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]504<i>; External declaration of the puts function</i>&nbsp;
505<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]506
507<i>; Definition of main function</i>
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]508define i32 @main() { <i>; i32()* </i>&nbsp;
509 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
510 %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]511
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]512 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
513 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
514 <a href="#i_ret">ret</a> i32 0&nbsp;
515}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]516
517<i>; Named metadata</i>
518!1 = metadata !{i32 41}
519!foo = !{!1, null}
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]520</pre>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]521
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]522<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]523 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]524 a <a href="#functionstructure">function definition</a> for
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]525 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
526 "<tt>foo"</tt>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]527
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]528<p>In general, a module is made up of a list of global values, where both
529 functions and global variables are global values. Global values are
530 represented by a pointer to a memory location (in this case, a pointer to an
531 array of char, and a pointer to a function), and have one of the
532 following <a href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]533
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]534</div>
535
536<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]537<h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]538 <a name="linkage">Linkage Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]539</h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]540
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]541<div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]542
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]543<p>All Global Variables and Functions have one of the following types of
544 linkage:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]545
546<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]547 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]548 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
549 by objects in the current module. In particular, linking code into a
550 module with an private global value may cause the private to be renamed as
551 necessary to avoid collisions. Because the symbol is private to the
552 module, all references can be updated. This doesn't show up in any symbol
553 table in the object file.</dd>
Rafael Espindolabb46f522009-01-15 20:18:42 +0000[diff] [blame]554
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]555 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]556 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
557 assembler and evaluated by the linker. Unlike normal strong symbols, they
558 are removed by the linker from the final linked image (executable or
559 dynamic library).</dd>
560
561 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
562 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
563 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
564 linker. The symbols are removed by the linker from the final linked image
565 (executable or dynamic library).</dd>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]566
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]567 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
568 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
569 of the object is not taken. For instance, functions that had an inline
570 definition, but the compiler decided not to inline it. Note,
571 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
572 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
573 visibility. The symbols are removed by the linker from the final linked
574 image (executable or dynamic library).</dd>
575
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]576 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling07d31772010-06-29 22:34:52 +0000[diff] [blame]577 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]578 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
579 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]580
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]581 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]582 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]583 into the object file corresponding to the LLVM module. They exist to
584 allow inlining and other optimizations to take place given knowledge of
585 the definition of the global, which is known to be somewhere outside the
586 module. Globals with <tt>available_externally</tt> linkage are allowed to
587 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
588 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]589
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]590 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]591 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner873187c2010-01-09 19:15:14 +0000[diff] [blame]592 the same name when linkage occurs. This can be used to implement
593 some forms of inline functions, templates, or other code which must be
594 generated in each translation unit that uses it, but where the body may
595 be overridden with a more definitive definition later. Unreferenced
596 <tt>linkonce</tt> globals are allowed to be discarded. Note that
597 <tt>linkonce</tt> linkage does not actually allow the optimizer to
598 inline the body of this function into callers because it doesn't know if
599 this definition of the function is the definitive definition within the
600 program or whether it will be overridden by a stronger definition.
601 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
602 linkage.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]603
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]604 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]605 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
606 <tt>linkonce</tt> linkage, except that unreferenced globals with
607 <tt>weak</tt> linkage may not be discarded. This is used for globals that
608 are declared "weak" in C source code.</dd>
609
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]610 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]611 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
612 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
613 global scope.
614 Symbols with "<tt>common</tt>" linkage are merged in the same way as
615 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]616 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]617 must have a zero initializer, and may not be marked '<a
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]618 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
619 have common linkage.</dd>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]620
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]621
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]622 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]623 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]624 pointer to array type. When two global variables with appending linkage
625 are linked together, the two global arrays are appended together. This is
626 the LLVM, typesafe, equivalent of having the system linker append together
627 "sections" with identical names when .o files are linked.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]628
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]629 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]630 <dd>The semantics of this linkage follow the ELF object file model: the symbol
631 is weak until linked, if not linked, the symbol becomes null instead of
632 being an undefined reference.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]633
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]634 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
635 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]636 <dd>Some languages allow differing globals to be merged, such as two functions
637 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]638 that only equivalent globals are ever merged (the "one definition rule"
639 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]640 and <tt>weak_odr</tt> linkage types to indicate that the global will only
641 be merged with equivalent globals. These linkage types are otherwise the
642 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]643
Bill Wendling5c3a9f72011-11-04 20:40:41 +0000[diff] [blame]644 <dt><tt><b><a name="linkage_external">external</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]645 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]646 visible, meaning that it participates in linkage and can be used to
647 resolve external symbol references.</dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]648</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]649
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]650<p>The next two types of linkage are targeted for Microsoft Windows platform
651 only. They are designed to support importing (exporting) symbols from (to)
652 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]653
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]654<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]655 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]656 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]657 or variable via a global pointer to a pointer that is set up by the DLL
658 exporting the symbol. On Microsoft Windows targets, the pointer name is
659 formed by combining <code>__imp_</code> and the function or variable
660 name.</dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]661
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]662 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]663 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]664 pointer to a pointer in a DLL, so that it can be referenced with the
665 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
666 name is formed by combining <code>__imp_</code> and the function or
667 variable name.</dd>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]668</dl>
669
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]670<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
671 another module defined a "<tt>.LC0</tt>" variable and was linked with this
672 one, one of the two would be renamed, preventing a collision. Since
673 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
674 declarations), they are accessible outside of the current module.</p>
675
676<p>It is illegal for a function <i>declaration</i> to have any linkage type
Bill Wendlingf7f06102011-10-11 06:41:28 +0000[diff] [blame]677 other than <tt>external</tt>, <tt>dllimport</tt>
678 or <tt>extern_weak</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]679
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]680<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]681 or <tt>weak_odr</tt> linkages.</p>
682
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]683</div>
684
685<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]686<h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]687 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]688</h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]689
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]690<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]691
692<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]693 and <a href="#i_invoke">invokes</a> can all have an optional calling
694 convention specified for the call. The calling convention of any pair of
695 dynamic caller/callee must match, or the behavior of the program is
696 undefined. The following calling conventions are supported by LLVM, and more
697 may be added in the future:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]698
699<dl>
700 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]701 <dd>This calling convention (the default if no other calling convention is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]702 specified) matches the target C calling conventions. This calling
703 convention supports varargs function calls and tolerates some mismatch in
704 the declared prototype and implemented declaration of the function (as
705 does normal C).</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]706
707 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]708 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]709 (e.g. by passing things in registers). This calling convention allows the
710 target to use whatever tricks it wants to produce fast code for the
711 target, without having to conform to an externally specified ABI
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]712 (Application Binary Interface).
713 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]714 when this or the GHC convention is used.</a> This calling convention
715 does not support varargs and requires the prototype of all callees to
716 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]717
718 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]719 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]720 as possible under the assumption that the call is not commonly executed.
721 As such, these calls often preserve all registers so that the call does
722 not break any live ranges in the caller side. This calling convention
723 does not support varargs and requires the prototype of all callees to
724 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]725
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]726 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
727 <dd>This calling convention has been implemented specifically for use by the
728 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
729 It passes everything in registers, going to extremes to achieve this by
730 disabling callee save registers. This calling convention should not be
731 used lightly but only for specific situations such as an alternative to
732 the <em>register pinning</em> performance technique often used when
733 implementing functional programming languages.At the moment only X86
734 supports this convention and it has the following limitations:
735 <ul>
736 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
737 floating point types are supported.</li>
738 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
739 6 floating point parameters.</li>
740 </ul>
741 This calling convention supports
742 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
743 requires both the caller and callee are using it.
744 </dd>
745
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]746 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]747 <dd>Any calling convention may be specified by number, allowing
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]748 target-specific calling conventions to be used. Target specific calling
749 conventions start at 64.</dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]750</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]751
752<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]753 support Pascal conventions or any other well-known target-independent
754 convention.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]755
756</div>
757
758<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]759<h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]760 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]761</h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]762
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]763<div>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]764
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]765<p>All Global Variables and Functions have one of the following visibility
766 styles:</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]767
768<dl>
769 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]770 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]771 that the declaration is visible to other modules and, in shared libraries,
772 means that the declared entity may be overridden. On Darwin, default
773 visibility means that the declaration is visible to other modules. Default
774 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]775
776 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]777 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]778 object if they are in the same shared object. Usually, hidden visibility
779 indicates that the symbol will not be placed into the dynamic symbol
780 table, so no other module (executable or shared library) can reference it
781 directly.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]782
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]783 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]784 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]785 the dynamic symbol table, but that references within the defining module
786 will bind to the local symbol. That is, the symbol cannot be overridden by
787 another module.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]788</dl>
789
790</div>
791
792<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]793<h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]794 <a name="namedtypes">Named Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]795</h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]796
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]797<div>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]798
799<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]800 it easier to read the IR and make the IR more condensed (particularly when
801 recursive types are involved). An example of a name specification is:</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]802
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]803<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]804%mytype = type { %mytype*, i32 }
805</pre>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]806
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]807<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattnerdc65f222010-08-17 23:26:04 +0000[diff] [blame]808 "<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]809 is expected with the syntax "%mytype".</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]810
811<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]812 and that you can therefore specify multiple names for the same type. This
813 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
814 uses structural typing, the name is not part of the type. When printing out
815 LLVM IR, the printer will pick <em>one name</em> to render all types of a
816 particular shape. This means that if you have code where two different
817 source types end up having the same LLVM type, that the dumper will sometimes
818 print the "wrong" or unexpected type. This is an important design point and
819 isn't going to change.</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]820
821</div>
822
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]823<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]824<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]825 <a name="globalvars">Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]826</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]827
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]828<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]829
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]830<p>Global variables define regions of memory allocated at compilation time
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]831 instead of run-time. Global variables may optionally be initialized, may
832 have an explicit section to be placed in, and may have an optional explicit
833 alignment specified. A variable may be defined as "thread_local", which
834 means that it will not be shared by threads (each thread will have a
835 separated copy of the variable). A variable may be defined as a global
836 "constant," which indicates that the contents of the variable
837 will <b>never</b> be modified (enabling better optimization, allowing the
838 global data to be placed in the read-only section of an executable, etc).
839 Note that variables that need runtime initialization cannot be marked
840 "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]841
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]842<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
843 constant, even if the final definition of the global is not. This capability
844 can be used to enable slightly better optimization of the program, but
845 requires the language definition to guarantee that optimizations based on the
846 'constantness' are valid for the translation units that do not include the
847 definition.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]848
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]849<p>As SSA values, global variables define pointer values that are in scope
850 (i.e. they dominate) all basic blocks in the program. Global variables
851 always define a pointer to their "content" type because they describe a
852 region of memory, and all memory objects in LLVM are accessed through
853 pointers.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]854
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]855<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
856 that the address is not significant, only the content. Constants marked
Rafael Espindolaa5eaa862011-01-15 08:20:57 +0000[diff] [blame]857 like this can be merged with other constants if they have the same
858 initializer. Note that a constant with significant address <em>can</em>
859 be merged with a <tt>unnamed_addr</tt> constant, the result being a
860 constant whose address is significant.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]861
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]862<p>A global variable may be declared to reside in a target-specific numbered
863 address space. For targets that support them, address spaces may affect how
864 optimizations are performed and/or what target instructions are used to
865 access the variable. The default address space is zero. The address space
866 qualifier must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000[diff] [blame]867
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]868<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]869 supports it, it will emit globals to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]870
Chris Lattnerce99fa92010-04-28 00:13:42 +0000[diff] [blame]871<p>An explicit alignment may be specified for a global, which must be a power
872 of 2. If not present, or if the alignment is set to zero, the alignment of
873 the global is set by the target to whatever it feels convenient. If an
874 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner2d4b8ee2010-04-28 00:31:12 +0000[diff] [blame]875 alignment. Targets and optimizers are not allowed to over-align the global
876 if the global has an assigned section. In this case, the extra alignment
877 could be observable: for example, code could assume that the globals are
878 densely packed in their section and try to iterate over them as an array,
879 alignment padding would break this iteration.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]880
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]881<p>For example, the following defines a global in a numbered address space with
882 an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]883
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]884<pre class="doc_code">
Dan Gohman398873c2009-01-11 00:40:00 +0000[diff] [blame]885@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]886</pre>
887
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]888</div>
889
890
891<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]892<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]893 <a name="functionstructure">Functions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]894</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]895
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]896<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]897
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]898<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]899 optional <a href="#linkage">linkage type</a>, an optional
900 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]901 <a href="#callingconv">calling convention</a>,
902 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]903 <a href="#paramattrs">parameter attribute</a> for the return type, a function
904 name, a (possibly empty) argument list (each with optional
905 <a href="#paramattrs">parameter attributes</a>), optional
906 <a href="#fnattrs">function attributes</a>, an optional section, an optional
907 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
908 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]909
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]910<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
911 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]912 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]913 <a href="#callingconv">calling convention</a>,
914 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]915 <a href="#paramattrs">parameter attribute</a> for the return type, a function
916 name, a possibly empty list of arguments, an optional alignment, and an
917 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]918
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]919<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]920 (Control Flow Graph) for the function. Each basic block may optionally start
921 with a label (giving the basic block a symbol table entry), contains a list
922 of instructions, and ends with a <a href="#terminators">terminator</a>
923 instruction (such as a branch or function return).</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]924
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]925<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]926 executed on entrance to the function, and it is not allowed to have
927 predecessor basic blocks (i.e. there can not be any branches to the entry
928 block of a function). Because the block can have no predecessors, it also
929 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]930
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]931<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]932 supports it, it will emit functions to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]933
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]934<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]935 the alignment is set to zero, the alignment of the function is set by the
936 target to whatever it feels convenient. If an explicit alignment is
937 specified, the function is forced to have at least that much alignment. All
938 alignments must be a power of 2.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]939
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]940<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
Bill Wendling5c3a9f72011-11-04 20:40:41 +0000[diff] [blame]941 be significant and two identical functions can be merged.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]942
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]943<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]944<pre class="doc_code">
Chris Lattner50ad45c2008-10-13 16:55:18 +0000[diff] [blame]945define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]946 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
947 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
948 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
949 [<a href="#gc">gc</a>] { ... }
950</pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]951
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]952</div>
953
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]954<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]955<h3>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]956 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]957</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]958
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]959<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]960
961<p>Aliases act as "second name" for the aliasee value (which can be either
962 function, global variable, another alias or bitcast of global value). Aliases
963 may have an optional <a href="#linkage">linkage type</a>, and an
964 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]965
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]966<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]967<pre class="doc_code">
Duncan Sands0b23ac12008-09-12 20:48:21 +0000[diff] [blame]968@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]969</pre>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]970
971</div>
972
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]973<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]974<h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]975 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]976</h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]977
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]978<div>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]979
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]980<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]981 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]982 a named metadata.</p>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]983
984<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]985<pre class="doc_code">
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]986; Some unnamed metadata nodes, which are referenced by the named metadata.
987!0 = metadata !{metadata !"zero"}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]988!1 = metadata !{metadata !"one"}
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]989!2 = metadata !{metadata !"two"}
Dan Gohman1005bc52010-07-13 19:48:13 +0000[diff] [blame]990; A named metadata.
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]991!name = !{!0, !1, !2}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]992</pre>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]993
994</div>
995
996<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]997<h3>
998 <a name="paramattrs">Parameter Attributes</a>
999</h3>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1000
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1001<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1002
1003<p>The return type and each parameter of a function type may have a set of
1004 <i>parameter attributes</i> associated with them. Parameter attributes are
1005 used to communicate additional information about the result or parameters of
1006 a function. Parameter attributes are considered to be part of the function,
1007 not of the function type, so functions with different parameter attributes
1008 can have the same function type.</p>
1009
1010<p>Parameter attributes are simple keywords that follow the type specified. If
1011 multiple parameter attributes are needed, they are space separated. For
1012 example:</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1013
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1014<pre class="doc_code">
Nick Lewyckyb6a7d252009-02-15 23:06:14 +0000[diff] [blame]1015declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner66d922c2008-10-04 18:33:34 +0000[diff] [blame]1016declare i32 @atoi(i8 zeroext)
1017declare signext i8 @returns_signed_char()
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1018</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1019
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1020<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1021 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1022
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1023<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1024
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1025<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1026 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1027 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichebe81732011-03-16 22:20:18 +0000[diff] [blame]1028 should be zero-extended to the extent required by the target's ABI (which
1029 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1030 parameter) or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1031
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1032 <dt><tt><b>signext</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 Zwarich9e69ff92011-03-17 14:21:58 +0000[diff] [blame]1034 should be sign-extended to the extent required by the target's ABI (which
1035 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1036 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>inreg</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1039 <dd>This indicates that this parameter or return value should be treated in a
1040 special target-dependent fashion during while emitting code for a function
1041 call or return (usually, by putting it in a register as opposed to memory,
1042 though some targets use it to distinguish between two different kinds of
1043 registers). Use of this attribute is target-specific.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1044
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1045 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1046 <dd><p>This indicates that the pointer parameter should really be passed by
1047 value to the function. The attribute implies that a hidden copy of the
1048 pointee
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1049 is made between the caller and the callee, so the callee is unable to
1050 modify the value in the callee. This attribute is only valid on LLVM
1051 pointer arguments. It is generally used to pass structs and arrays by
1052 value, but is also valid on pointers to scalars. The copy is considered
1053 to belong to the caller not the callee (for example,
1054 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1055 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1056 values.</p>
1057
1058 <p>The byval attribute also supports specifying an alignment with
1059 the align attribute. It indicates the alignment of the stack slot to
1060 form and the known alignment of the pointer specified to the call site. If
1061 the alignment is not specified, then the code generator makes a
1062 target-specific assumption.</p></dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1063
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1064 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1065 <dd>This indicates that the pointer parameter specifies the address of a
1066 structure that is the return value of the function in the source program.
1067 This pointer must be guaranteed by the caller to be valid: loads and
1068 stores to the structure may be assumed by the callee to not to trap. This
1069 may only be applied to the first parameter. This is not a valid attribute
1070 for return values. </dd>
1071
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1072 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1073 <dd>This indicates that pointer values
1074 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmanefca7f92010-07-02 23:46:54 +0000[diff] [blame]1075 value do not alias pointer values which are not <i>based</i> on it,
1076 ignoring certain "irrelevant" dependencies.
1077 For a call to the parent function, dependencies between memory
1078 references from before or after the call and from those during the call
1079 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1080 return value used in that call.
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1081 The caller shares the responsibility with the callee for ensuring that
1082 these requirements are met.
1083 For further details, please see the discussion of the NoAlias response in
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1084 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1085<br>
John McCall191d4ee2010-07-06 21:07:14 +0000[diff] [blame]1086 Note that this definition of <tt>noalias</tt> is intentionally
1087 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner211244a2010-07-06 20:51:35 +0000[diff] [blame]1088 arguments, though it is slightly weaker.
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1089<br>
1090 For function return values, C99's <tt>restrict</tt> is not meaningful,
1091 while LLVM's <tt>noalias</tt> is.
1092 </dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1093
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1094 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1095 <dd>This indicates that the callee does not make any copies of the pointer
1096 that outlive the callee itself. This is not a valid attribute for return
1097 values.</dd>
1098
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1099 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1100 <dd>This indicates that the pointer parameter can be excised using the
1101 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1102 attribute for return values.</dd>
1103</dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1104
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1105</div>
1106
1107<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1108<h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1109 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1110</h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1111
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1112<div>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1113
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1114<p>Each function may specify a garbage collector name, which is simply a
1115 string:</p>
1116
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1117<pre class="doc_code">
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1118define void @f() gc "name" { ... }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1119</pre>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1120
1121<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1122 collector which will cause the compiler to alter its output in order to
1123 support the named garbage collection algorithm.</p>
1124
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1125</div>
1126
1127<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1128<h3>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1129 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1130</h3>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1131
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1132<div>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1133
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1134<p>Function attributes are set to communicate additional information about a
1135 function. Function attributes are considered to be part of the function, not
1136 of the function type, so functions with different parameter attributes can
1137 have the same function type.</p>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1138
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1139<p>Function attributes are simple keywords that follow the type specified. If
1140 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1141
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1142<pre class="doc_code">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1143define void @f() noinline { ... }
1144define void @f() alwaysinline { ... }
1145define void @f() alwaysinline optsize { ... }
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1146define void @f() optsize { ... }
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1147</pre>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1148
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1149<dl>
Kostya Serebryany164b86b2012-01-20 17:56:17 +0000[diff] [blame]1150 <dt><tt><b>address_safety</b></tt></dt>
1151 <dd>This attribute indicates that the address safety analysis
1152 is enabled for this function. </dd>
1153
Charles Davis1e063d12010-02-12 00:31:15 +0000[diff] [blame]1154 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1155 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1156 the backend should forcibly align the stack pointer. Specify the
1157 desired alignment, which must be a power of two, in parentheses.
1158
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1159 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1160 <dd>This attribute indicates that the inliner should attempt to inline this
1161 function into callers whenever possible, ignoring any active inlining size
1162 threshold for this caller.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1163
Dan Gohman129bd562011-06-16 16:03:13 +0000[diff] [blame]1164 <dt><tt><b>nonlazybind</b></tt></dt>
1165 <dd>This attribute suppresses lazy symbol binding for the function. This
1166 may make calls to the function faster, at the cost of extra program
1167 startup time if the function is not called during program startup.</dd>
1168
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +0000[diff] [blame]1169 <dt><tt><b>inlinehint</b></tt></dt>
1170 <dd>This attribute indicates that the source code contained a hint that inlining
1171 this function is desirable (such as the "inline" keyword in C/C++). It
1172 is just a hint; it imposes no requirements on the inliner.</dd>
1173
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1174 <dt><tt><b>naked</b></tt></dt>
1175 <dd>This attribute disables prologue / epilogue emission for the function.
1176 This can have very system-specific consequences.</dd>
1177
1178 <dt><tt><b>noimplicitfloat</b></tt></dt>
1179 <dd>This attributes disables implicit floating point instructions.</dd>
1180
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1181 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1182 <dd>This attribute indicates that the inliner should never inline this
1183 function in any situation. This attribute may not be used together with
1184 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1185
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1186 <dt><tt><b>noredzone</b></tt></dt>
1187 <dd>This attribute indicates that the code generator should not use a red
1188 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1189
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1190 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1191 <dd>This function attribute indicates that the function never returns
1192 normally. This produces undefined behavior at runtime if the function
1193 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +0000[diff] [blame]1194
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1195 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1196 <dd>This function attribute indicates that the function never returns with an
1197 unwind or exceptional control flow. If the function does unwind, its
1198 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +0000[diff] [blame]1199
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1200 <dt><tt><b>optsize</b></tt></dt>
1201 <dd>This attribute suggests that optimization passes and code generator passes
1202 make choices that keep the code size of this function low, and otherwise
1203 do optimizations specifically to reduce code size.</dd>
1204
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1205 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1206 <dd>This attribute indicates that the function computes its result (or decides
1207 to unwind an exception) based strictly on its arguments, without
1208 dereferencing any pointer arguments or otherwise accessing any mutable
1209 state (e.g. memory, control registers, etc) visible to caller functions.
1210 It does not write through any pointer arguments
1211 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1212 changes any state visible to callers. This means that it cannot unwind
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]1213 exceptions by calling the <tt>C++</tt> exception throwing methods.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +0000[diff] [blame]1214
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1215 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1216 <dd>This attribute indicates that the function does not write through any
1217 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1218 arguments) or otherwise modify any state (e.g. memory, control registers,
1219 etc) visible to caller functions. It may dereference pointer arguments
1220 and read state that may be set in the caller. A readonly function always
1221 returns the same value (or unwinds an exception identically) when called
1222 with the same set of arguments and global state. It cannot unwind an
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]1223 exception by calling the <tt>C++</tt> exception throwing methods.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +0000[diff] [blame]1224
Bill Wendling9bd5d042011-12-05 21:27:54 +0000[diff] [blame]1225 <dt><tt><b><a name="returns_twice">returns_twice</a></b></tt></dt>
1226 <dd>This attribute indicates that this function can return twice. The
1227 C <code>setjmp</code> is an example of such a function. The compiler
1228 disables some optimizations (like tail calls) in the caller of these
1229 functions.</dd>
1230
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1231 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1232 <dd>This attribute indicates that the function should emit a stack smashing
1233 protector. It is in the form of a "canary"&mdash;a random value placed on
1234 the stack before the local variables that's checked upon return from the
1235 function to see if it has been overwritten. A heuristic is used to
1236 determine if a function needs stack protectors or not.<br>
1237<br>
1238 If a function that has an <tt>ssp</tt> attribute is inlined into a
1239 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1240 function will have an <tt>ssp</tt> attribute.</dd>
1241
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1242 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1243 <dd>This attribute indicates that the function should <em>always</em> emit a
1244 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1245 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1246<br>
1247 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1248 function that doesn't have an <tt>sspreq</tt> attribute or which has
1249 an <tt>ssp</tt> attribute, then the resulting function will have
1250 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindolafbff0ec2011-07-25 15:27:59 +0000[diff] [blame]1251
1252 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1253 <dd>This attribute indicates that the ABI being targeted requires that
1254 an unwind table entry be produce for this function even if we can
1255 show that no exceptions passes by it. This is normally the case for
1256 the ELF x86-64 abi, but it can be disabled for some compilation
1257 units.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1258</dl>
1259
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1260</div>
1261
1262<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1263<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]1264 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1265</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1266
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1267<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1268
1269<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1270 the GCC "file scope inline asm" blocks. These blocks are internally
1271 concatenated by LLVM and treated as a single unit, but may be separated in
1272 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1273
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1274<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1275module asm "inline asm code goes here"
1276module asm "more can go here"
1277</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1278
1279<p>The strings can contain any character by escaping non-printable characters.
1280 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]1281 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1282
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1283<p>The inline asm code is simply printed to the machine code .s file when
1284 assembly code is generated.</p>
1285
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1286</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1287
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1288<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1289<h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1290 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1291</h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1292
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1293<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1294
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1295<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1296 data is to be laid out in memory. The syntax for the data layout is
1297 simply:</p>
1298
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1299<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1300target datalayout = "<i>layout specification</i>"
1301</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1302
1303<p>The <i>layout specification</i> consists of a list of specifications
1304 separated by the minus sign character ('-'). Each specification starts with
1305 a letter and may include other information after the letter to define some
1306 aspect of the data layout. The specifications accepted are as follows:</p>
1307
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1308<dl>
1309 <dt><tt>E</tt></dt>
1310 <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]1311 bits with the most significance have the lowest address location.</dd>
1312
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1313 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1314 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1315 the bits with the least significance have the lowest address
1316 location.</dd>
1317
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1318 <dt><tt>S<i>size</i></tt></dt>
1319 <dd>Specifies the natural alignment of the stack in bits. Alignment promotion
1320 of stack variables is limited to the natural stack alignment to avoid
1321 dynamic stack realignment. The stack alignment must be a multiple of
Lang Hames5f119a62011-10-11 17:50:14 +0000[diff] [blame]1322 8-bits. If omitted, the natural stack alignment defaults to "unspecified",
1323 which does not prevent any alignment promotions.</dd>
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1324
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1325 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1326 <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]1327 <i>preferred</i> alignments. All sizes are in bits. Specifying
1328 the <i>pref</i> alignment is optional. If omitted, the
1329 preceding <tt>:</tt> should be omitted too.</dd>
1330
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1331 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1332 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1333 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1334
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1335 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1336 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1337 <i>size</i>.</dd>
1338
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1339 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1340 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +0000[diff] [blame]1341 <i>size</i>. Only values of <i>size</i> that are supported by the target
1342 will work. 32 (float) and 64 (double) are supported on all targets;
1343 80 or 128 (different flavors of long double) are also supported on some
1344 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1345
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1346 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1347 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1348 <i>size</i>.</dd>
1349
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1350 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1351 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1352 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1353
1354 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1355 <dd>This specifies a set of native integer widths for the target CPU
1356 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1357 "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]1358 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1359 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1360</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1361
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1362<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]1363 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1364 specifications in the <tt>datalayout</tt> keyword. The default specifications
1365 are given in this list:</p>
1366
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1367<ul>
1368 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +0000[diff] [blame]1369 <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]1370 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1371 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1372 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1373 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1374 <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]1375 alignment of 64-bits</li>
1376 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1377 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1378 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1379 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1380 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1381 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1382</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1383
1384<p>When LLVM is determining the alignment for a given type, it uses the
1385 following rules:</p>
1386
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1387<ol>
1388 <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]1389 specification is used.</li>
1390
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1391 <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]1392 smallest integer type that is larger than the bitwidth of the sought type
1393 is used. If none of the specifications are larger than the bitwidth then
1394 the the largest integer type is used. For example, given the default
1395 specifications above, the i7 type will use the alignment of i8 (next
1396 largest) while both i65 and i256 will use the alignment of i64 (largest
1397 specified).</li>
1398
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1399 <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]1400 largest vector type that is smaller than the sought vector type will be
1401 used as a fall back. This happens because &lt;128 x double&gt; can be
1402 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1403</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1404
Chris Lattner6509f502011-10-11 23:01:39 +0000[diff] [blame]1405<p>The function of the data layout string may not be what you expect. Notably,
1406 this is not a specification from the frontend of what alignment the code
1407 generator should use.</p>
1408
1409<p>Instead, if specified, the target data layout is required to match what the
1410 ultimate <em>code generator</em> expects. This string is used by the
1411 mid-level optimizers to
1412 improve code, and this only works if it matches what the ultimate code
1413 generator uses. If you would like to generate IR that does not embed this
1414 target-specific detail into the IR, then you don't have to specify the
1415 string. This will disable some optimizations that require precise layout
1416 information, but this also prevents those optimizations from introducing
1417 target specificity into the IR.</p>
1418
1419
1420
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1421</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1422
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1423<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1424<h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1425 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1426</h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1427
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1428<div>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1429
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1430<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +0000[diff] [blame]1431with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1432is undefined. Pointer values are associated with address ranges
1433according to the following rules:</p>
1434
1435<ul>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1436 <li>A pointer value is associated with the addresses associated with
1437 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1438 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1439 range of the variable's storage.</li>
1440 <li>The result value of an allocation instruction is associated with
1441 the address range of the allocated storage.</li>
1442 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1443 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1444 <li>An integer constant other than zero or a pointer value returned
1445 from a function not defined within LLVM may be associated with address
1446 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1447 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1448 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1449</ul>
1450
1451<p>A pointer value is <i>based</i> on another pointer value according
1452 to the following rules:</p>
1453
1454<ul>
1455 <li>A pointer value formed from a
1456 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1457 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1458 <li>The result value of a
1459 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1460 of the <tt>bitcast</tt>.</li>
1461 <li>A pointer value formed by an
1462 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1463 pointer values that contribute (directly or indirectly) to the
1464 computation of the pointer's value.</li>
1465 <li>The "<i>based</i> on" relationship is transitive.</li>
1466</ul>
1467
1468<p>Note that this definition of <i>"based"</i> is intentionally
1469 similar to the definition of <i>"based"</i> in C99, though it is
1470 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1471
1472<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1473<tt><a href="#i_load">load</a></tt> merely indicates the size and
1474alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +0000[diff] [blame]1475interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1476<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1477and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1478
1479<p>Consequently, type-based alias analysis, aka TBAA, aka
1480<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1481LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1482additional information which specialized optimization passes may use
1483to implement type-based alias analysis.</p>
1484
1485</div>
1486
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1487<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1488<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1489 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1490</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1491
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1492<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1493
1494<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1495href="#i_store"><tt>store</tt></a>s, and <a
1496href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1497The optimizers must not change the number of volatile operations or change their
1498order of execution relative to other volatile operations. The optimizers
1499<i>may</i> change the order of volatile operations relative to non-volatile
1500operations. This is not Java's "volatile" and has no cross-thread
1501synchronization behavior.</p>
1502
1503</div>
1504
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1505<!-- ======================================================================= -->
1506<h3>
1507 <a name="memmodel">Memory Model for Concurrent Operations</a>
1508</h3>
1509
1510<div>
1511
1512<p>The LLVM IR does not define any way to start parallel threads of execution
1513or to register signal handlers. Nonetheless, there are platform-specific
1514ways to create them, and we define LLVM IR's behavior in their presence. This
1515model is inspired by the C++0x memory model.</p>
1516
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1517<p>For a more informal introduction to this model, see the
1518<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.
1519
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1520<p>We define a <i>happens-before</i> partial order as the least partial order
1521that</p>
1522<ul>
1523 <li>Is a superset of single-thread program order, and</li>
1524 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1525 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1526 by platform-specific techniques, like pthread locks, thread
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1527 creation, thread joining, etc., and by atomic instructions.
1528 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1529 </li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1530</ul>
1531
1532<p>Note that program order does not introduce <i>happens-before</i> edges
1533between a thread and signals executing inside that thread.</p>
1534
1535<p>Every (defined) read operation (load instructions, memcpy, atomic
1536loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1537(defined) write operations (store instructions, atomic
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1538stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1539initialized globals are considered to have a write of the initializer which is
1540atomic and happens before any other read or write of the memory in question.
1541For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1542any write to the same byte, except:</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1543
1544<ul>
1545 <li>If <var>write<sub>1</sub></var> happens before
1546 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1547 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1548 does not see <var>write<sub>1</sub></var>.
Bill Wendling0246bb72011-07-31 06:45:03 +0000[diff] [blame]1549 <li>If <var>R<sub>byte</sub></var> happens before
1550 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1551 see <var>write<sub>3</sub></var>.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1552</ul>
1553
1554<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1555<ul>
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1556 <li>If <var>R</var> is volatile, the result is target-dependent. (Volatile
1557 is supposed to give guarantees which can support
1558 <code>sig_atomic_t</code> in C/C++, and may be used for accesses to
1559 addresses which do not behave like normal memory. It does not generally
1560 provide cross-thread synchronization.)
1561 <li>Otherwise, if there is no write to the same byte that happens before
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1562 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1563 <tt>undef</tt> for that byte.
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1564 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1565 <var>R<sub>byte</sub></var> returns the value written by that
1566 write.</li>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1567 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1568 <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]1569 values written. See the <a href="#ordering">Atomic Memory Ordering
1570 Constraints</a> section for additional constraints on how the choice
1571 is made.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1572 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1573</ul>
1574
1575<p><var>R</var> returns the value composed of the series of bytes it read.
1576This implies that some bytes within the value may be <tt>undef</tt>
1577<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1578defines the semantics of the operation; it doesn't mean that targets will
1579emit more than one instruction to read the series of bytes.</p>
1580
1581<p>Note that in cases where none of the atomic intrinsics are used, this model
1582places only one restriction on IR transformations on top of what is required
1583for single-threaded execution: introducing a store to a byte which might not
Eli Friedman101c81d2011-08-02 01:15:34 +0000[diff] [blame]1584otherwise be stored is not allowed in general. (Specifically, in the case
1585where another thread might write to and read from an address, introducing a
1586store can change a load that may see exactly one write into a load that may
1587see multiple writes.)</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1588
1589<!-- FIXME: This model assumes all targets where concurrency is relevant have
1590a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1591none of the backends currently in the tree fall into this category; however,
1592there might be targets which care. If there are, we want a paragraph
1593like the following:
1594
1595Targets may specify that stores narrower than a certain width are not
1596available; on such a target, for the purposes of this model, treat any
1597non-atomic write with an alignment or width less than the minimum width
1598as if it writes to the relevant surrounding bytes.
1599-->
1600
1601</div>
1602
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1603<!-- ======================================================================= -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1604<h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1605 <a name="ordering">Atomic Memory Ordering Constraints</a>
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1606</h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1607
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1608<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1609
1610<p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]1611<a href="#i_atomicrmw"><code>atomicrmw</code></a>,
1612<a href="#i_fence"><code>fence</code></a>,
1613<a href="#i_load"><code>atomic load</code></a>, and
Eli Friedman8fa281a2011-08-09 23:26:12 +0000[diff] [blame]1614<a href="#i_store"><code>atomic store</code></a>) take an ordering parameter
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1615that determines which other atomic instructions on the same address they
1616<i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1617but are somewhat more colloquial. If these descriptions aren't precise enough,
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1618check those specs (see spec references in the
Nick Lewycky300a2632012-01-23 08:47:21 +0000[diff] [blame]1619<a href="Atomics.html#introduction">atomics guide</a>).
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1620<a href="#i_fence"><code>fence</code></a> instructions
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1621treat these orderings somewhat differently since they don't take an address.
1622See that instruction's documentation for details.</p>
1623
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1624<p>For a simpler introduction to the ordering constraints, see the
1625<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.</p>
1626
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1627<dl>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1628<dt><code>unordered</code></dt>
1629<dd>The set of values that can be read is governed by the happens-before
1630partial order. A value cannot be read unless some operation wrote it.
1631This is intended to provide a guarantee strong enough to model Java's
1632non-volatile shared variables. This ordering cannot be specified for
1633read-modify-write operations; it is not strong enough to make them atomic
1634in any interesting way.</dd>
1635<dt><code>monotonic</code></dt>
1636<dd>In addition to the guarantees of <code>unordered</code>, there is a single
1637total order for modifications by <code>monotonic</code> operations on each
1638address. All modification orders must be compatible with the happens-before
1639order. There is no guarantee that the modification orders can be combined to
1640a global total order for the whole program (and this often will not be
1641possible). The read in an atomic read-modify-write operation
1642(<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1643<a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1644reads the value in the modification order immediately before the value it
1645writes. If one atomic read happens before another atomic read of the same
1646address, the later read must see the same value or a later value in the
1647address's modification order. This disallows reordering of
1648<code>monotonic</code> (or stronger) operations on the same address. If an
1649address is written <code>monotonic</code>ally by one thread, and other threads
1650<code>monotonic</code>ally read that address repeatedly, the other threads must
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1651eventually see the write. This corresponds to the C++0x/C1x
1652<code>memory_order_relaxed</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1653<dt><code>acquire</code></dt>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1654<dd>In addition to the guarantees of <code>monotonic</code>,
Eli Friedmanc264b2f2011-08-24 20:28:39 +0000[diff] [blame]1655a <i>synchronizes-with</i> edge may be formed with a <code>release</code>
1656operation. This is intended to model C++'s <code>memory_order_acquire</code>.</dd>
1657<dt><code>release</code></dt>
1658<dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1659writes a value which is subsequently read by an <code>acquire</code> operation,
1660it <i>synchronizes-with</i> that operation. (This isn't a complete
1661description; see the C++0x definition of a release sequence.) This corresponds
1662to the C++0x/C1x <code>memory_order_release</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1663<dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1664<code>acquire</code> and <code>release</code> operation on its address.
1665This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1666<dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1667<dd>In addition to the guarantees of <code>acq_rel</code>
1668(<code>acquire</code> for an operation which only reads, <code>release</code>
1669for an operation which only writes), there is a global total order on all
1670sequentially-consistent operations on all addresses, which is consistent with
1671the <i>happens-before</i> partial order and with the modification orders of
1672all the affected addresses. Each sequentially-consistent read sees the last
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1673preceding write to the same address in this global order. This corresponds
1674to the C++0x/C1x <code>memory_order_seq_cst</code> and Java volatile.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1675</dl>
1676
1677<p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1678it only <i>synchronizes with</i> or participates in modification and seq_cst
1679total orderings with other operations running in the same thread (for example,
1680in signal handlers).</p>
1681
1682</div>
1683
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1684</div>
1685
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1686<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1687<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1688<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1689
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1690<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1691
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1692<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1693 intermediate representation. Being typed enables a number of optimizations
1694 to be performed on the intermediate representation directly, without having
1695 to do extra analyses on the side before the transformation. A strong type
1696 system makes it easier to read the generated code and enables novel analyses
1697 and transformations that are not feasible to perform on normal three address
1698 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1699
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1700<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1701<h3>
1702 <a name="t_classifications">Type Classifications</a>
1703</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1704
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1705<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1706
1707<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1708
1709<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1710 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1711 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1712 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1713 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1714 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1715 </tr>
1716 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1717 <td><a href="#t_floating">floating point</a></td>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1718 <td><tt>half, float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1719 </tr>
1720 <tr>
1721 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1722 <td><a href="#t_integer">integer</a>,
1723 <a href="#t_floating">floating point</a>,
1724 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +0000[diff] [blame]1725 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]1726 <a href="#t_struct">structure</a>,
1727 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1728 <a href="#t_label">label</a>,
1729 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1730 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1731 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1732 <tr>
1733 <td><a href="#t_primitive">primitive</a></td>
1734 <td><a href="#t_label">label</a>,
1735 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +0000[diff] [blame]1736 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1737 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1738 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1739 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1740 </tr>
1741 <tr>
1742 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1743 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1744 <a href="#t_function">function</a>,
1745 <a href="#t_pointer">pointer</a>,
1746 <a href="#t_struct">structure</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1747 <a href="#t_vector">vector</a>,
1748 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +0000[diff] [blame]1749 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1750 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1751 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1752</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1753
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1754<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1755 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1756 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1757
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1758</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1759
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1760<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1761<h3>
1762 <a name="t_primitive">Primitive Types</a>
1763</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1764
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1765<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1766
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1767<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1768 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1769
1770<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1771<h4>
1772 <a name="t_integer">Integer Type</a>
1773</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1774
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1775<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1776
1777<h5>Overview:</h5>
1778<p>The integer type is a very simple type that simply specifies an arbitrary
1779 bit width for the integer type desired. Any bit width from 1 bit to
1780 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1781
1782<h5>Syntax:</h5>
1783<pre>
1784 iN
1785</pre>
1786
1787<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1788 value.</p>
1789
1790<h5>Examples:</h5>
1791<table class="layout">
1792 <tr class="layout">
1793 <td class="left"><tt>i1</tt></td>
1794 <td class="left">a single-bit integer.</td>
1795 </tr>
1796 <tr class="layout">
1797 <td class="left"><tt>i32</tt></td>
1798 <td class="left">a 32-bit integer.</td>
1799 </tr>
1800 <tr class="layout">
1801 <td class="left"><tt>i1942652</tt></td>
1802 <td class="left">a really big integer of over 1 million bits.</td>
1803 </tr>
1804</table>
1805
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1806</div>
1807
1808<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1809<h4>
1810 <a name="t_floating">Floating Point Types</a>
1811</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1812
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1813<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1814
1815<table>
1816 <tbody>
1817 <tr><th>Type</th><th>Description</th></tr>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1818 <tr><td><tt>half</tt></td><td>16-bit floating point value</td></tr>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1819 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1820 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1821 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1822 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1823 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1824 </tbody>
1825</table>
1826
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1827</div>
1828
1829<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1830<h4>
1831 <a name="t_x86mmx">X86mmx Type</a>
1832</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1833
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1834<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1835
1836<h5>Overview:</h5>
1837<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>
1838
1839<h5>Syntax:</h5>
1840<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +0000[diff] [blame]1841 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1842</pre>
1843
1844</div>
1845
1846<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1847<h4>
1848 <a name="t_void">Void Type</a>
1849</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1850
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1851<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1852
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1853<h5>Overview:</h5>
1854<p>The void type does not represent any value and has no size.</p>
1855
1856<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1857<pre>
1858 void
1859</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1860
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1861</div>
1862
1863<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1864<h4>
1865 <a name="t_label">Label Type</a>
1866</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1867
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1868<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1869
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1870<h5>Overview:</h5>
1871<p>The label type represents code labels.</p>
1872
1873<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1874<pre>
1875 label
1876</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1877
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1878</div>
1879
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1880<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1881<h4>
1882 <a name="t_metadata">Metadata Type</a>
1883</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1884
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1885<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1886
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1887<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1888<p>The metadata type represents embedded metadata. No derived types may be
1889 created from metadata except for <a href="#t_function">function</a>
1890 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1891
1892<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1893<pre>
1894 metadata
1895</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1896
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1897</div>
1898
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1899</div>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1900
1901<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1902<h3>
1903 <a name="t_derived">Derived Types</a>
1904</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1905
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1906<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1907
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1908<p>The real power in LLVM comes from the derived types in the system. This is
1909 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1910 useful types. Each of these types contain one or more element types which
1911 may be a primitive type, or another derived type. For example, it is
1912 possible to have a two dimensional array, using an array as the element type
1913 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1914
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1915<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1916<h4>
1917 <a name="t_aggregate">Aggregate Types</a>
1918</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1919
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1920<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1921
1922<p>Aggregate Types are a subset of derived types that can contain multiple
Duncan Sands20536b52011-12-14 15:44:20 +0000[diff] [blame]1923 member types. <a href="#t_array">Arrays</a> and
1924 <a href="#t_struct">structs</a> are aggregate types.
1925 <a href="#t_vector">Vectors</a> are not considered to be aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1926
1927</div>
1928
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1929<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1930<h4>
1931 <a name="t_array">Array Type</a>
1932</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1933
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1934<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1935
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1936<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1937<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1938 sequentially in memory. The array type requires a size (number of elements)
1939 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1940
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1941<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1942<pre>
1943 [&lt;# elements&gt; x &lt;elementtype&gt;]
1944</pre>
1945
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1946<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1947 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1948
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1949<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1950<table class="layout">
1951 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1952 <td class="left"><tt>[40 x i32]</tt></td>
1953 <td class="left">Array of 40 32-bit integer values.</td>
1954 </tr>
1955 <tr class="layout">
1956 <td class="left"><tt>[41 x i32]</tt></td>
1957 <td class="left">Array of 41 32-bit integer values.</td>
1958 </tr>
1959 <tr class="layout">
1960 <td class="left"><tt>[4 x i8]</tt></td>
1961 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1962 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1963</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1964<p>Here are some examples of multidimensional arrays:</p>
1965<table class="layout">
1966 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1967 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1968 <td class="left">3x4 array of 32-bit integer values.</td>
1969 </tr>
1970 <tr class="layout">
1971 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1972 <td class="left">12x10 array of single precision floating point values.</td>
1973 </tr>
1974 <tr class="layout">
1975 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1976 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1977 </tr>
1978</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1979
Dan Gohman7657f6b2009-11-09 19:01:53 +0000[diff] [blame]1980<p>There is no restriction on indexing beyond the end of the array implied by
1981 a static type (though there are restrictions on indexing beyond the bounds
1982 of an allocated object in some cases). This means that single-dimension
1983 'variable sized array' addressing can be implemented in LLVM with a zero
1984 length array type. An implementation of 'pascal style arrays' in LLVM could
1985 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1986
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1987</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1988
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1989<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1990<h4>
1991 <a name="t_function">Function Type</a>
1992</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1993
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1994<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1995
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1996<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1997<p>The function type can be thought of as a function signature. It consists of
1998 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]1999 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]2000
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2001<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2002<pre>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2003 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2004</pre>
2005
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2006<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]2007 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
2008 which indicates that the function takes a variable number of arguments.
2009 Variable argument functions can access their arguments with
2010 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2011 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]2012 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2013
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2014<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2015<table class="layout">
2016 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2017 <td class="left"><tt>i32 (i32)</tt></td>
2018 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2019 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2020 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2021 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]2022 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2023 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2024 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
2025 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2026 </td>
2027 </tr><tr class="layout">
2028 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2029 <td class="left">A vararg function that takes at least one
2030 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2031 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2032 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2033 </td>
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2034 </tr><tr class="layout">
2035 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2036 <td class="left">A function taking an <tt>i32</tt>, returning a
2037 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2038 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2039 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2040</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2041
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2042</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2043
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2044<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2045<h4>
2046 <a name="t_struct">Structure Type</a>
2047</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2048
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2049<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2050
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2051<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2052<p>The structure type is used to represent a collection of data members together
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2053 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]2054
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]2055<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2056 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2057 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2058 Structures in registers are accessed using the
2059 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2060 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2061
2062<p>Structures may optionally be "packed" structures, which indicate that the
2063 alignment of the struct is one byte, and that there is no padding between
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2064 the elements. In non-packed structs, padding between field types is inserted
2065 as defined by the TargetData string in the module, which is required to match
Chris Lattnere4617b02011-10-11 23:02:17 +0000[diff] [blame]2066 what the underlying code generator expects.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2067
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2068<p>Structures can either be "literal" or "identified". A literal structure is
2069 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) whereas identified
2070 types are always defined at the top level with a name. Literal types are
2071 uniqued by their contents and can never be recursive or opaque since there is
Chris Lattneraa175c32011-08-12 18:12:40 +0000[diff] [blame]2072 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]2073 never uniqued.
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2074</p>
2075
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2076<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2077<pre>
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2078 %T1 = type { &lt;type list&gt; } <i>; Identified normal struct type</i>
2079 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Identified packed struct type</i>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2080</pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2081
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2082<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2083<table class="layout">
2084 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2085 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2086 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2087 </tr>
2088 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2089 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
2090 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2091 second element is a <a href="#t_pointer">pointer</a> to a
2092 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2093 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2094 </tr>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2095 <tr class="layout">
2096 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
2097 <td class="left">A packed struct known to be 5 bytes in size.</td>
2098 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2099</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2100
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2101</div>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2102
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2103<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2104<h4>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2105 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2106</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2107
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2108<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2109
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2110<h5>Overview:</h5>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2111<p>Opaque structure types are used to represent named structure types that do
2112 not have a body specified. This corresponds (for example) to the C notion of
2113 a forward declared structure.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2114
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2115<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2116<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2117 %X = type opaque
2118 %52 = type opaque
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2119</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2120
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2121<h5>Examples:</h5>
2122<table class="layout">
2123 <tr class="layout">
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2124 <td class="left"><tt>opaque</tt></td>
2125 <td class="left">An opaque type.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2126 </tr>
2127</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2128
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2129</div>
2130
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2131
2132
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2133<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2134<h4>
2135 <a name="t_pointer">Pointer Type</a>
2136</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2137
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2138<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2139
2140<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +0000[diff] [blame]2141<p>The pointer type is used to specify memory locations.
2142 Pointers are commonly used to reference objects in memory.</p>
2143
2144<p>Pointer types may have an optional address space attribute defining the
2145 numbered address space where the pointed-to object resides. The default
2146 address space is number zero. The semantics of non-zero address
2147 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2148
2149<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2150 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2151
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2152<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2153<pre>
2154 &lt;type&gt; *
2155</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2156
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2157<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2158<table class="layout">
2159 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2160 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2161 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2162 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2163 </tr>
2164 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]2165 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2166 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2167 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2168 <tt>i32</tt>.</td>
2169 </tr>
2170 <tr class="layout">
2171 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2172 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2173 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2174 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2175</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2176
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2177</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2178
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2179<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2180<h4>
2181 <a name="t_vector">Vector Type</a>
2182</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2183
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2184<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2185
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2186<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2187<p>A vector type is a simple derived type that represents a vector of elements.
2188 Vector types are used when multiple primitive data are operated in parallel
2189 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +0000[diff] [blame]2190 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2191 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2192
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2193<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2194<pre>
2195 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2196</pre>
2197
Chris Lattner7d2e7be2010-10-10 18:20:35 +0000[diff] [blame]2198<p>The number of elements is a constant integer value larger than 0; elementtype
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2199 may be any integer or floating point type, or a pointer to these types.
2200 Vectors of size zero are not allowed. </p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2201
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2202<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2203<table class="layout">
2204 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2205 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2206 <td class="left">Vector of 4 32-bit integer values.</td>
2207 </tr>
2208 <tr class="layout">
2209 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2210 <td class="left">Vector of 8 32-bit floating-point values.</td>
2211 </tr>
2212 <tr class="layout">
2213 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2214 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2215 </tr>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2216 <tr class="layout">
2217 <td class="left"><tt>&lt;4 x i64*&gt;</tt></td>
2218 <td class="left">Vector of 4 pointers to 64-bit integer values.</td>
2219 </tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2220</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2221
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2222</div>
2223
Bill Wendlingaf75f0c2011-07-31 06:47:33 +0000[diff] [blame]2224</div>
2225
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]2226</div>
2227
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2228<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2229<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2230<!-- *********************************************************************** -->
2231
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2232<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2233
2234<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2235 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2236
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2237<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2238<h3>
2239 <a name="simpleconstants">Simple Constants</a>
2240</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2241
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2242<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2243
2244<dl>
2245 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2246 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]2247 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2248
2249 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2250 <dd>Standard integers (such as '4') are constants of
2251 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2252 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2253
2254 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2255 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2256 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2257 notation (see below). The assembler requires the exact decimal value of a
2258 floating-point constant. For example, the assembler accepts 1.25 but
2259 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2260 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2261
2262 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2263 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2264 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2265</dl>
2266
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2267<p>The one non-intuitive notation for constants is the hexadecimal form of
2268 floating point constants. For example, the form '<tt>double
2269 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2270 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2271 constants are required (and the only time that they are generated by the
2272 disassembler) is when a floating point constant must be emitted but it cannot
2273 be represented as a decimal floating point number in a reasonable number of
2274 digits. For example, NaN's, infinities, and other special values are
2275 represented in their IEEE hexadecimal format so that assembly and disassembly
2276 do not cause any bits to change in the constants.</p>
2277
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2278<p>When using the hexadecimal form, constants of types half, float, and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2279 represented using the 16-digit form shown above (which matches the IEEE754
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2280 representation for double); half and float values must, however, be exactly
2281 representable as IEE754 half and single precision, respectively.
2282 Hexadecimal format is always used
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2283 for long double, and there are three forms of long double. The 80-bit format
2284 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2285 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2286 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2287 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2288 currently supported target uses this format. Long doubles will only work if
2289 they match the long double format on your target. All hexadecimal formats
2290 are big-endian (sign bit at the left).</p>
2291
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]2292<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2293</div>
2294
2295<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2296<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]2297<a name="aggregateconstants"></a> <!-- old anchor -->
2298<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2299</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2300
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2301<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2302
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2303<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2304 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2305
2306<dl>
2307 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2308 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2309 type definitions (a comma separated list of elements, surrounded by braces
2310 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2311 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2312 Structure constants must have <a href="#t_struct">structure type</a>, and
2313 the number and types of elements must match those specified by the
2314 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2315
2316 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2317 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2318 definitions (a comma separated list of elements, surrounded by square
2319 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2320 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2321 the number and types of elements must match those specified by the
2322 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2323
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2324 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2325 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2326 definitions (a comma separated list of elements, surrounded by
2327 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2328 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2329 have <a href="#t_vector">vector type</a>, and the number and types of
2330 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2331
2332 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2333 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]2334 value to zero of <em>any</em> type, including scalar and
2335 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2336 This is often used to avoid having to print large zero initializers
2337 (e.g. for large arrays) and is always exactly equivalent to using explicit
2338 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +0000[diff] [blame]2339
2340 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +0000[diff] [blame]2341 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2342 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2343 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2344 be interpreted as part of the instruction stream, metadata is a place to
2345 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2346</dl>
2347
2348</div>
2349
2350<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2351<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2352 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2353</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2354
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2355<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2356
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2357<p>The addresses of <a href="#globalvars">global variables</a>
2358 and <a href="#functionstructure">functions</a> are always implicitly valid
2359 (link-time) constants. These constants are explicitly referenced when
2360 the <a href="#identifiers">identifier for the global</a> is used and always
2361 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2362 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2363
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2364<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]2365@X = global i32 17
2366@Y = global i32 42
2367@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2368</pre>
2369
2370</div>
2371
2372<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2373<h3>
2374 <a name="undefvalues">Undefined Values</a>
2375</h3>
2376
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2377<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2378
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2379<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]2380 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2381 Undefined values may be of any type (other than '<tt>label</tt>'
2382 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2383
Chris Lattnerc608cb12009-09-11 01:49:31 +0000[diff] [blame]2384<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2385 program is well defined no matter what value is used. This gives the
2386 compiler more freedom to optimize. Here are some examples of (potentially
2387 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2388
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2389
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2390<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2391 %A = add %X, undef
2392 %B = sub %X, undef
2393 %C = xor %X, undef
2394Safe:
2395 %A = undef
2396 %B = undef
2397 %C = undef
2398</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2399
2400<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]2401 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]2402
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2403<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2404 %A = or %X, undef
2405 %B = and %X, undef
2406Safe:
2407 %A = -1
2408 %B = 0
2409Unsafe:
2410 %A = undef
2411 %B = undef
2412</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2413
2414<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2415 For example, if <tt>%X</tt> has a zero bit, then the output of the
2416 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2417 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2418 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2419 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2420 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2421 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2422 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2423
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2424<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2425 %A = select undef, %X, %Y
2426 %B = select undef, 42, %Y
2427 %C = select %X, %Y, undef
2428Safe:
2429 %A = %X (or %Y)
2430 %B = 42 (or %Y)
2431 %C = %Y
2432Unsafe:
2433 %A = undef
2434 %B = undef
2435 %C = undef
2436</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2437
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2438<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2439 branch) conditions can go <em>either way</em>, but they have to come from one
2440 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2441 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2442 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2443 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2444 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2445 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2446
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2447<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2448 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2449
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2450 %B = undef
2451 %C = xor %B, %B
2452
2453 %D = undef
2454 %E = icmp lt %D, 4
2455 %F = icmp gte %D, 4
2456
2457Safe:
2458 %A = undef
2459 %B = undef
2460 %C = undef
2461 %D = undef
2462 %E = undef
2463 %F = undef
2464</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2465
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2466<p>This example points out that two '<tt>undef</tt>' operands are not
2467 necessarily the same. This can be surprising to people (and also matches C
2468 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2469 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2470 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2471 its value over its "live range". This is true because the variable doesn't
2472 actually <em>have a live range</em>. Instead, the value is logically read
2473 from arbitrary registers that happen to be around when needed, so the value
2474 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2475 need to have the same semantics or the core LLVM "replace all uses with"
2476 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2477
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2478<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2479 %A = fdiv undef, %X
2480 %B = fdiv %X, undef
2481Safe:
2482 %A = undef
2483b: unreachable
2484</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2485
2486<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2487 value</em> and <em>undefined behavior</em>. An undefined value (like
2488 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2489 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2490 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2491 defined on SNaN's. However, in the second example, we can make a more
2492 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2493 arbitrary value, we are allowed to assume that it could be zero. Since a
2494 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2495 the operation does not execute at all. This allows us to delete the divide and
2496 all code after it. Because the undefined operation "can't happen", the
2497 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2498
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2499<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2500a: store undef -> %X
2501b: store %X -> undef
2502Safe:
2503a: &lt;deleted&gt;
2504b: unreachable
2505</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2506
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2507<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2508 undefined value can be assumed to not have any effect; we can assume that the
2509 value is overwritten with bits that happen to match what was already there.
2510 However, a store <em>to</em> an undefined location could clobber arbitrary
2511 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2512
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2513</div>
2514
2515<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2516<h3>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2517 <a name="poisonvalues">Poison Values</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2518</h3>
2519
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2520<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2521
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2522<p>Poison values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2523 they also represent the fact that an instruction or constant expression which
2524 cannot evoke side effects has nevertheless detected a condition which results
2525 in undefined behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2526
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2527<p>There is currently no way of representing a poison value in the IR; they
Dan Gohman855abed2010-05-03 14:51:43 +0000[diff] [blame]2528 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2529 <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]2530
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2531<p>Poison value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2532
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2533<ul>
2534<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2535 their operands.</li>
2536
2537<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2538 to their dynamic predecessor basic block.</li>
2539
2540<li>Function arguments depend on the corresponding actual argument values in
2541 the dynamic callers of their functions.</li>
2542
2543<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2544 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2545 control back to them.</li>
2546
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2547<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]2548 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_resume"><tt>resume</tt></a>,
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2549 or exception-throwing call instructions that dynamically transfer control
2550 back to them.</li>
2551
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2552<li>Non-volatile loads and stores depend on the most recent stores to all of the
2553 referenced memory addresses, following the order in the IR
2554 (including loads and stores implied by intrinsics such as
2555 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2556
Dan Gohman7c24ff12010-05-03 14:59:34 +0000[diff] [blame]2557<!-- TODO: In the case of multiple threads, this only applies if the store
2558 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2559
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2560<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2561
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2562<li>An instruction with externally visible side effects depends on the most
2563 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]2564 the order in the IR. (This includes
2565 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2566
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2567<li>An instruction <i>control-depends</i> on a
2568 <a href="#terminators">terminator instruction</a>
2569 if the terminator instruction has multiple successors and the instruction
2570 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2571 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2572
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2573<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2574 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2575 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2576 successor.</li>
2577
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2578<li>Dependence is transitive.</li>
2579
2580</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2581
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2582<p>Poison Values have the same behavior as <a href="#undefvalues">undef values</a>,
2583 with the additional affect that any instruction which has a <i>dependence</i>
2584 on a poison value has undefined behavior.</p>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2585
2586<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +0000[diff] [blame]2587
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2588<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2589entry:
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2590 %poison = sub nuw i32 0, 1 ; Results in a poison value.
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2591 %still_poison = and i32 %poison, 0 ; 0, but also poison.
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2592 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2593 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2594
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2595 store i32 %poison, i32* @g ; Poison value stored to memory.
2596 %poison2 = load i32* @g ; Poison value loaded back from memory.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2597
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2598 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2599
2600 %narrowaddr = bitcast i32* @g to i16*
2601 %wideaddr = bitcast i32* @g to i64*
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2602 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2603 %poison4 = load i64* %wideaddr ; Returns a poison value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2604
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2605 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2606 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2607
2608true:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2609 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2610 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2611 br label %end
2612
2613end:
2614 %p = phi i32 [ 0, %entry ], [ 1, %true ]
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2615 ; Both edges into this PHI are
2616 ; control-dependent on %cmp, so this
2617 ; always results in a poison value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2618
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2619 store volatile i32 0, i32* @g ; This would depend on the store in %true
2620 ; if %cmp is true, or the store in %entry
2621 ; otherwise, so this is undefined behavior.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2622
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2623 br i1 %cmp, label %second_true, label %second_end
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2624 ; The same branch again, but this time the
2625 ; true block doesn't have side effects.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2626
2627second_true:
2628 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2629 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2630
2631second_end:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2632 store volatile i32 0, i32* @g ; This time, the instruction always depends
2633 ; on the store in %end. Also, it is
2634 ; control-equivalent to %end, so this is
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2635 ; well-defined (ignoring earlier undefined
2636 ; behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2637</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2638
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2639</div>
2640
2641<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2642<h3>
2643 <a name="blockaddress">Addresses of Basic Blocks</a>
2644</h3>
2645
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2646<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2647
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2648<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2649
2650<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +0000[diff] [blame]2651 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2652 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2653
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2654<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2655 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2656 comparisons against null. Pointer equality tests between labels addresses
2657 results in undefined behavior &mdash; though, again, comparison against null
2658 is ok, and no label is equal to the null pointer. This may be passed around
2659 as an opaque pointer sized value as long as the bits are not inspected. This
2660 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2661 long as the original value is reconstituted before the <tt>indirectbr</tt>
2662 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2663
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2664<p>Finally, some targets may provide defined semantics when using the value as
2665 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2666
2667</div>
2668
2669
2670<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2671<h3>
2672 <a name="constantexprs">Constant Expressions</a>
2673</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2674
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2675<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2676
2677<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2678 to be used as constants. Constant expressions may be of
2679 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2680 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2681 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2682
2683<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2684 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2685 <dd>Truncate a constant to another type. The bit size of CST must be larger
2686 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2687
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2688 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2689 <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]2690 smaller than the bit size of TYPE. Both types must be integers.</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>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2693 <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]2694 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2695
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2696 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2697 <dd>Truncate a floating point constant to another floating point type. The
2698 size of CST must be larger than the size of TYPE. Both types must be
2699 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2700
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2701 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2702 <dd>Floating point extend a constant to another type. The size of CST must be
2703 smaller or equal to the size of TYPE. Both types must be floating
2704 point.</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>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2707 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2708 constant. TYPE must be a scalar or vector integer type. CST must be of
2709 scalar or vector floating point type. Both CST and TYPE must be scalars,
2710 or vectors of the same number of elements. If the value won't fit in the
2711 integer 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>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2714 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2715 constant. TYPE must be a scalar or vector integer type. CST must be of
2716 scalar or vector floating point type. Both CST and TYPE must be scalars,
2717 or vectors of the same number of elements. If the value won't fit in the
2718 integer 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>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2721 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2722 constant. TYPE must be a scalar or vector floating point type. CST must be
2723 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2724 vectors of the same number of elements. If the value won't fit in the
2725 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2726
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2727 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2728 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2729 constant. TYPE must be a scalar or vector floating point type. CST must be
2730 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2731 vectors of the same number of elements. If the value won't fit in the
2732 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2733
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2734 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2735 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2736 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2737 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2738 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2739
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2740 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2741 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2742 type. CST must be of integer type. The CST value is zero extended,
2743 truncated, or unchanged to make it fit in a pointer size. This one is
2744 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2745
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2746 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +0000[diff] [blame]2747 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2748 are the same as those for the <a href="#i_bitcast">bitcast
2749 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2750
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2751 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2752 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2753 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2754 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2755 instruction, the index list may have zero or more indexes, which are
2756 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2757
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2758 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2759 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2760
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2761 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2762 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2763
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2764 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2765 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2766
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2767 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2768 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2769 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2770
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2771 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2772 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2773 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2774
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2775 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2776 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2777 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2778
Nick Lewycky9e130ce2010-05-29 06:44:15 +0000[diff] [blame]2779 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2780 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2781 constants. The index list is interpreted in a similar manner as indices in
2782 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2783 index value must be specified.</dd>
2784
2785 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2786 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2787 constants. The index list is interpreted in a similar manner as indices in
2788 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2789 index value must be specified.</dd>
2790
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2791 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2792 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2793 be any of the <a href="#binaryops">binary</a>
2794 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2795 on operands are the same as those for the corresponding instruction
2796 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2797</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2798
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2799</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]2800
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2801</div>
2802
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2803<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2804<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2805<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2806<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2807<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2808<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2809<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2810</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2811
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2812<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2813
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2814<p>LLVM supports inline assembler expressions (as opposed
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2815 to <a href="#moduleasm">Module-Level Inline Assembly</a>) through the use of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2816 a special value. This value represents the inline assembler as a string
2817 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2818 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2819 expression has side effects, and a flag indicating whether the function
2820 containing the asm needs to align its stack conservatively. An example
2821 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2822
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2823<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2824i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2825</pre>
2826
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2827<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2828 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2829 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2830
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2831<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2832%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]2833</pre>
2834
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2835<p>Inline asms with side effects not visible in the constraint list must be
2836 marked as having side effects. This is done through the use of the
2837 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2838
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2839<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2840call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2841</pre>
2842
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2843<p>In some cases inline asms will contain code that will not work unless the
2844 stack is aligned in some way, such as calls or SSE instructions on x86,
2845 yet will not contain code that does that alignment within the asm.
2846 The compiler should make conservative assumptions about what the asm might
2847 contain and should generate its usual stack alignment code in the prologue
2848 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2849
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2850<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2851call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2852</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2853
2854<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2855 first.</p>
2856
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2857<!--
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2858<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2859 documented here. Constraints on what can be done (e.g. duplication, moving,
2860 etc need to be documented). This is probably best done by reference to
2861 another document that covers inline asm from a holistic perspective.</p>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2862 -->
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2863
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2864<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2865<h4>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2866 <a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2867</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2868
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2869<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2870
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2871<p>The call instructions that wrap inline asm nodes may have a
2872 "<tt>!srcloc</tt>" MDNode attached to it that contains a list of constant
2873 integers. If present, the code generator will use the integer as the
2874 location cookie value when report errors through the <tt>LLVMContext</tt>
2875 error reporting mechanisms. This allows a front-end to correlate backend
2876 errors that occur with inline asm back to the source code that produced it.
2877 For example:</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2878
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2879<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2880call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2881...
2882!42 = !{ i32 1234567 }
2883</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2884
2885<p>It is up to the front-end to make sense of the magic numbers it places in the
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2886 IR. If the MDNode contains multiple constants, the code generator will use
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2887 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]2888
2889</div>
2890
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2891</div>
2892
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2893<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2894<h3>
2895 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2896</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2897
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2898<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2899
2900<p>LLVM IR allows metadata to be attached to instructions in the program that
2901 can convey extra information about the code to the optimizers and code
2902 generator. One example application of metadata is source-level debug
2903 information. There are two metadata primitives: strings and nodes. All
2904 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2905 preceding exclamation point ('<tt>!</tt>').</p>
2906
2907<p>A metadata string is a string surrounded by double quotes. It can contain
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2908 any character by escaping non-printable characters with "<tt>\xx</tt>" where
2909 "<tt>xx</tt>" is the two digit hex code. For example:
2910 "<tt>!"test\00"</tt>".</p>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2911
2912<p>Metadata nodes are represented with notation similar to structure constants
2913 (a comma separated list of elements, surrounded by braces and preceded by an
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2914 exclamation point). Metadata nodes can have any values as their operand. For
2915 example:</p>
2916
2917<div class="doc_code">
2918<pre>
2919!{ metadata !"test\00", i32 10}
2920</pre>
2921</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2922
2923<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2924 metadata nodes, which can be looked up in the module symbol table. For
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2925 example:</p>
2926
2927<div class="doc_code">
2928<pre>
2929!foo = metadata !{!4, !3}
2930</pre>
2931</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2932
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2933<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2934 function is using two metadata arguments:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2935
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2936<div class="doc_code">
2937<pre>
2938call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2939</pre>
2940</div>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2941
2942<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2943 attached to the <tt>add</tt> instruction using the <tt>!dbg</tt>
2944 identifier:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2945
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2946<div class="doc_code">
2947<pre>
2948%indvar.next = add i64 %indvar, 1, !dbg !21
2949</pre>
2950</div>
2951
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]2952<p>More information about specific metadata nodes recognized by the optimizers
2953 and code generator is found below.</p>
2954
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2955<!-- _______________________________________________________________________ -->
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]2956<h4>
2957 <a name="tbaa">'<tt>tbaa</tt>' Metadata</a>
2958</h4>
2959
2960<div>
2961
2962<p>In LLVM IR, memory does not have types, so LLVM's own type system is not
2963 suitable for doing TBAA. Instead, metadata is added to the IR to describe
2964 a type system of a higher level language. This can be used to implement
2965 typical C/C++ TBAA, but it can also be used to implement custom alias
2966 analysis behavior for other languages.</p>
2967
2968<p>The current metadata format is very simple. TBAA metadata nodes have up to
2969 three fields, e.g.:</p>
2970
2971<div class="doc_code">
2972<pre>
2973!0 = metadata !{ metadata !"an example type tree" }
2974!1 = metadata !{ metadata !"int", metadata !0 }
2975!2 = metadata !{ metadata !"float", metadata !0 }
2976!3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2977</pre>
2978</div>
2979
2980<p>The first field is an identity field. It can be any value, usually
2981 a metadata string, which uniquely identifies the type. The most important
2982 name in the tree is the name of the root node. Two trees with
2983 different root node names are entirely disjoint, even if they
2984 have leaves with common names.</p>
2985
2986<p>The second field identifies the type's parent node in the tree, or
2987 is null or omitted for a root node. A type is considered to alias
2988 all of its descendants and all of its ancestors in the tree. Also,
2989 a type is considered to alias all types in other trees, so that
2990 bitcode produced from multiple front-ends is handled conservatively.</p>
2991
2992<p>If the third field is present, it's an integer which if equal to 1
2993 indicates that the type is "constant" (meaning
2994 <tt>pointsToConstantMemory</tt> should return true; see
2995 <a href="AliasAnalysis.html#OtherItfs">other useful
2996 <tt>AliasAnalysis</tt> methods</a>).</p>
2997
2998</div>
2999
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]3000<!-- _______________________________________________________________________ -->
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3001<h4>
3002 <a name="fpaccuracy">'<tt>fpaccuracy</tt>' Metadata</a>
3003</h4>
3004
3005<div>
3006
3007<p><tt>fpaccuracy</tt> metadata may be attached to any instruction of floating
3008 point type. It expresses the maximum relative error of the result of
3009 that instruction, in ULPs. ULP is defined as follows:</p>
3010
Bill Wendling0656e252011-11-09 19:33:56 +0000[diff] [blame]3011<blockquote>
3012
3013<p>If <tt>x</tt> is a real number that lies between two finite consecutive
3014 floating-point numbers <tt>a</tt> and <tt>b</tt>, without being equal to one
3015 of them, then <tt>ulp(x) = |b - a|</tt>, otherwise <tt>ulp(x)</tt> is the
3016 distance between the two non-equal finite floating-point numbers nearest
3017 <tt>x</tt>. Moreover, <tt>ulp(NaN)</tt> is <tt>NaN</tt>.</p>
3018
3019</blockquote>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3020
3021<p>The maximum relative error may be any rational number. The metadata node
3022 shall consist of a pair of unsigned integers respectively representing
3023 the numerator and denominator. For example, 2.5 ULP:</p>
3024
3025<div class="doc_code">
3026<pre>
3027!0 = metadata !{ i32 5, i32 2 }
3028</pre>
3029</div>
3030
3031</div>
3032
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]3033</div>
3034
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3035</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3036
3037<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3038<h2>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3039 <a name="module_flags">Module Flags Metadata</a>
3040</h2>
3041<!-- *********************************************************************** -->
3042
3043<div>
3044
3045<p>Information about the module as a whole is difficult to convey to LLVM's
3046 subsystems. The LLVM IR isn't sufficient to transmit this
3047 information. The <tt>llvm.module.flags</tt> named metadata exists in order to
3048 facilitate this. These flags are in the form of key / value pairs &mdash;
3049 much like a dictionary &mdash; making it easy for any subsystem who cares
3050 about a flag to look it up.</p>
3051
3052<p>The <tt>llvm.module.flags</tt> metadata contains a list of metadata
3053 triplets. Each triplet has the following form:</p>
3054
3055<ul>
3056 <li>The first element is a <i>behavior</i> flag, which specifies the behavior
3057 when two (or more) modules are merged together, and it encounters two (or
3058 more) metadata with the same ID. The supported behaviors are described
3059 below.</li>
3060
3061 <li>The second element is a metadata string that is a unique ID for the
3062 metadata. How each ID is interpreted is documented below.</li>
3063
3064 <li>The third element is the value of the flag.</li>
3065</ul>
3066
3067<p>When two (or more) modules are merged together, the resulting
3068 <tt>llvm.module.flags</tt> metadata is the union of the
3069 modules' <tt>llvm.module.flags</tt> metadata. The only exception being a flag
3070 with the <i>Override</i> behavior, which may override another flag's value
3071 (see below).</p>
3072
3073<p>The following behaviors are supported:</p>
3074
3075<table border="1" cellspacing="0" cellpadding="4">
3076 <tbody>
3077 <tr>
3078 <th>Value</th>
3079 <th>Behavior</th>
3080 </tr>
3081 <tr>
3082 <td>1</td>
3083 <td align="left">
3084 <dt><b>Error</b></dt>
3085 <dd>Emits an error if two values disagree. It is an error to have an ID
3086 with both an Error and a Warning behavior.</dd>
3087 </td>
3088 </tr>
3089 <tr>
3090 <td>2</td>
3091 <td align="left">
3092 <dt><b>Warning</b></dt>
3093 <dd>Emits a warning if two values disagree.</dd>
3094 </td>
3095 </tr>
3096 <tr>
3097 <td>3</td>
3098 <td align="left">
3099 <dt><b>Require</b></dt>
3100 <dd>Emits an error when the specified value is not present or doesn't
3101 have the specified value. It is an error for two (or more)
3102 <tt>llvm.module.flags</tt> with the same ID to have the Require
3103 behavior but different values. There may be multiple Require flags
3104 per ID.</dd>
3105 </td>
3106 </tr>
3107 <tr>
3108 <td>4</td>
3109 <td align="left">
3110 <dt><b>Override</b></dt>
3111 <dd>Uses the specified value if the two values disagree. It is an error
3112 for two (or more) <tt>llvm.module.flags</tt> with the same ID to
3113 have the Override behavior but different values.</dd>
3114 </td>
3115 </tr>
3116 </tbody>
3117</table>
3118
3119<p>An example of module flags:</p>
3120
3121<pre class="doc_code">
3122!0 = metadata !{ i32 1, metadata !"foo", i32 1 }
3123!1 = metadata !{ i32 4, metadata !"bar", i32 37 }
3124!2 = metadata !{ i32 2, metadata !"qux", i32 42 }
3125!3 = metadata !{ i32 3, metadata !"qux",
3126 metadata !{
3127 metadata !"foo", i32 1
3128 }
3129}
3130!llvm.module.flags = !{ !0, !1, !2, !3 }
3131</pre>
3132
3133<ul>
3134 <li><p>Metadata <tt>!0</tt> has the ID <tt>!"foo"</tt> and the value '1'. The
3135 behavior if two or more <tt>!"foo"</tt> flags are seen is to emit an
3136 error if their values are not equal.</p></li>
3137
3138 <li><p>Metadata <tt>!1</tt> has the ID <tt>!"bar"</tt> and the value '37'. The
3139 behavior if two or more <tt>!"bar"</tt> flags are seen is to use the
3140 value '37' if their values are not equal.</p></li>
3141
3142 <li><p>Metadata <tt>!2</tt> has the ID <tt>!"qux"</tt> and the value '42'. The
3143 behavior if two or more <tt>!"qux"</tt> flags are seen is to emit a
3144 warning if their values are not equal.</p></li>
3145
3146 <li><p>Metadata <tt>!3</tt> has the ID <tt>!"qux"</tt> and the value:</p>
3147
3148<pre class="doc_code">
3149metadata !{ metadata !"foo", i32 1 }
3150</pre>
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3151
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3152 <p>The behavior is to emit an error if the <tt>llvm.module.flags</tt> does
3153 not contain a flag with the ID <tt>!"foo"</tt> that has the value
3154 '1'. If two or more <tt>!"qux"</tt> flags exist, then they must have
3155 the same value or an error will be issued.</p></li>
3156</ul>
3157
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3158
3159<!-- ======================================================================= -->
3160<h3>
3161<a name="objc_gc_flags">Objective-C Garbage Collection Module Flags Metadata</a>
3162</h3>
3163
3164<div>
3165
3166<p>On the Mach-O platform, Objective-C stores metadata about garbage collection
3167 in a special section called "image info". The metadata consists of a version
3168 number and a bitmask specifying what types of garbage collection are
3169 supported (if any) by the file. If two or more modules are linked together
3170 their garbage collection metadata needs to be merged rather than appended
3171 together.</p>
3172
3173<p>The Objective-C garbage collection module flags metadata consists of the
3174 following key-value pairs:</p>
3175
3176<table border="1" cellspacing="0" cellpadding="4">
3177 <tbody>
3178 <tr>
3179 <th width="30%">Key</th>
3180 <th>Value</th>
3181 </tr>
3182 <tr>
3183 <td><tt>Objective-C&nbsp;Version</tt></td>
3184 <td align="left"><b>[Required]</b> &mdash; The Objective-C ABI
3185 version. Valid values are 1 and 2.</td>
3186 </tr>
3187 <tr>
3188 <td><tt>Objective-C&nbsp;Image&nbsp;Info&nbsp;Version</tt></td>
3189 <td align="left"><b>[Required]</b> &mdash; The version of the image info
3190 section. Currently always 0.</td>
3191 </tr>
3192 <tr>
3193 <td><tt>Objective-C&nbsp;Image&nbsp;Info&nbsp;Section</tt></td>
3194 <td align="left"><b>[Required]</b> &mdash; The section to place the
3195 metadata. Valid values are <tt>"__OBJC, __image_info, regular"</tt> for
3196 Objective-C ABI version 1, and <tt>"__DATA,__objc_imageinfo, regular,
3197 no_dead_strip"</tt> for Objective-C ABI version 2.</td>
3198 </tr>
3199 <tr>
3200 <td><tt>Objective-C&nbsp;Garbage&nbsp;Collection</tt></td>
3201 <td align="left"><b>[Required]</b> &mdash; Specifies whether garbage
3202 collection is supported or not. Valid values are 0, for no garbage
3203 collection, and 2, for garbage collection supported.</td>
3204 </tr>
3205 <tr>
3206 <td><tt>Objective-C&nbsp;GC&nbsp;Only</tt></td>
3207 <td align="left"><b>[Optional]</b> &mdash; Specifies that only garbage
3208 collection is supported. If present, its value must be 6. This flag
3209 requires that the <tt>Objective-C Garbage Collection</tt> flag have the
3210 value 2.</td>
3211 </tr>
3212 </tbody>
3213</table>
3214
3215<p>Some important flag interactions:</p>
3216
3217<ul>
3218 <li>If a module with <tt>Objective-C Garbage Collection</tt> set to 0 is
3219 merged with a module with <tt>Objective-C Garbage Collection</tt> set to
3220 2, then the resulting module has the <tt>Objective-C Garbage
3221 Collection</tt> flag set to 0.</li>
3222
3223 <li>A module with <tt>Objective-C Garbage Collection</tt> set to 0 cannot be
3224 merged with a module with <tt>Objective-C GC Only</tt> set to 6.</li>
3225</ul>
3226
3227</div>
3228
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3229</div>
3230
3231<!-- *********************************************************************** -->
3232<h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3233 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3234</h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3235<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3236<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3237<p>LLVM has a number of "magic" global variables that contain data that affect
3238code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3239of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
3240section and all globals that start with "<tt>llvm.</tt>" are reserved for use
3241by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3242
3243<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3244<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3245<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3246</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3247
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3248<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3249
3250<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
3251href="#linkage_appending">appending linkage</a>. This array contains a list of
3252pointers to global variables and functions which may optionally have a pointer
3253cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
3254
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3255<div class="doc_code">
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3256<pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3257@X = global i8 4
3258@Y = global i32 123
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3259
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3260@llvm.used = appending global [2 x i8*] [
3261 i8* @X,
3262 i8* bitcast (i32* @Y to i8*)
3263], section "llvm.metadata"
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3264</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3265</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3266
3267<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3268 compiler, assembler, and linker are required to treat the symbol as if there
3269 is a reference to the global that it cannot see. For example, if a variable
3270 has internal linkage and no references other than that from
3271 the <tt>@llvm.used</tt> list, it cannot be deleted. This is commonly used to
3272 represent references from inline asms and other things the compiler cannot
3273 "see", and corresponds to "<tt>attribute((used))</tt>" in GNU C.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3274
3275<p>On some targets, the code generator must emit a directive to the assembler or
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3276 object file to prevent the assembler and linker from molesting the
3277 symbol.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3278
3279</div>
3280
3281<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3282<h3>
3283 <a name="intg_compiler_used">
3284 The '<tt>llvm.compiler.used</tt>' Global Variable
3285 </a>
3286</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3287
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3288<div>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3289
3290<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3291 <tt>@llvm.used</tt> directive, except that it only prevents the compiler from
3292 touching the symbol. On targets that support it, this allows an intelligent
3293 linker to optimize references to the symbol without being impeded as it would
3294 be by <tt>@llvm.used</tt>.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3295
3296<p>This is a rare construct that should only be used in rare circumstances, and
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3297 should not be exposed to source languages.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3298
3299</div>
3300
3301<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3302<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3303<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3304</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3305
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3306<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3307
3308<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3309<pre>
3310%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3311@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3312</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3313</div>
3314
3315<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor
3316 functions and associated priorities. The functions referenced by this array
3317 will be called in ascending order of priority (i.e. lowest first) when the
3318 module is loaded. The order of functions with the same priority is not
3319 defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3320
3321</div>
3322
3323<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3324<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3325<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3326</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3327
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3328<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3329
3330<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3331<pre>
3332%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3333@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3334</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3335</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3336
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3337<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions
3338 and associated priorities. The functions referenced by this array will be
3339 called in descending order of priority (i.e. highest first) when the module
3340 is loaded. The order of functions with the same priority is not defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3341
3342</div>
3343
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3344</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3345
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]3346<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3347<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3348<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3349
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3350<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3351
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3352<p>The LLVM instruction set consists of several different classifications of
3353 instructions: <a href="#terminators">terminator
3354 instructions</a>, <a href="#binaryops">binary instructions</a>,
3355 <a href="#bitwiseops">bitwise binary instructions</a>,
3356 <a href="#memoryops">memory instructions</a>, and
3357 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3358
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3359<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3360<h3>
3361 <a name="terminators">Terminator Instructions</a>
3362</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3363
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3364<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3365
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3366<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3367 in a program ends with a "Terminator" instruction, which indicates which
3368 block should be executed after the current block is finished. These
3369 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3370 control flow, not values (the one exception being the
3371 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3372
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3373<p>The terminator instructions are:
3374 '<a href="#i_ret"><tt>ret</tt></a>',
3375 '<a href="#i_br"><tt>br</tt></a>',
3376 '<a href="#i_switch"><tt>switch</tt></a>',
3377 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>',
3378 '<a href="#i_invoke"><tt>invoke</tt></a>',
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3379 '<a href="#i_resume"><tt>resume</tt></a>', and
3380 '<a href="#i_unreachable"><tt>unreachable</tt></a>'.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3381
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3382<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3383<h4>
3384 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3385</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3386
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3387<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3388
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3389<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3390<pre>
3391 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]3392 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3393</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3394
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3395<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3396<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3397 a value) from a function back to the caller.</p>
3398
3399<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3400 value and then causes control flow, and one that just causes control flow to
3401 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3402
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3403<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3404<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3405 return value. The type of the return value must be a
3406 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3407
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3408<p>A function is not <a href="#wellformed">well formed</a> if it it has a
3409 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3410 value or a return value with a type that does not match its type, or if it
3411 has a void return type and contains a '<tt>ret</tt>' instruction with a
3412 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3413
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3414<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3415<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3416 the calling function's context. If the caller is a
3417 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3418 instruction after the call. If the caller was an
3419 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3420 the beginning of the "normal" destination block. If the instruction returns
3421 a value, that value shall set the call or invoke instruction's return
3422 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3423
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3424<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3425<pre>
3426 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]3427 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +0000[diff] [blame]3428 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]3429</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]3430
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3431</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3432<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3433<h4>
3434 <a name="i_br">'<tt>br</tt>' Instruction</a>
3435</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3436
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3437<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3438
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3439<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3440<pre>
Bill Wendlingb3aa4712011-07-26 10:41:15 +0000[diff] [blame]3441 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3442 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3443</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3444
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3445<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3446<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3447 different basic block in the current function. There are two forms of this
3448 instruction, corresponding to a conditional branch and an unconditional
3449 branch.</p>
3450
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3451<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3452<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3453 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3454 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3455 target.</p>
3456
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3457<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3458<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3459 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3460 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3461 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3462
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3463<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]3464<pre>
3465Test:
3466 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3467 br i1 %cond, label %IfEqual, label %IfUnequal
3468IfEqual:
3469 <a href="#i_ret">ret</a> i32 1
3470IfUnequal:
3471 <a href="#i_ret">ret</a> i32 0
3472</pre>
3473
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3474</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3475
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3476<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3477<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3478 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3479</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3480
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3481<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3482
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3483<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3484<pre>
3485 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3486</pre>
3487
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3488<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3489<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]3490 several different places. It is a generalization of the '<tt>br</tt>'
3491 instruction, allowing a branch to occur to one of many possible
3492 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3493
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3494<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3495<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3496 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3497 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3498 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3499
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3500<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3501<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3502 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3503 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3504 transferred to the corresponding destination; otherwise, control flow is
3505 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3506
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3507<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3508<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3509 <tt>switch</tt> instruction, this instruction may be code generated in
3510 different ways. For example, it could be generated as a series of chained
3511 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3512
3513<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3514<pre>
3515 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3516 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3517 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3518
3519 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3520 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3521
3522 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3523 switch i32 %val, label %otherwise [ i32 0, label %onzero
3524 i32 1, label %onone
3525 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3526</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3527
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3528</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3529
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3530
3531<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3532<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3533 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3534</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3535
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3536<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3537
3538<h5>Syntax:</h5>
3539<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3540 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]3541</pre>
3542
3543<h5>Overview:</h5>
3544
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3545<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3546 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]3547 "<tt>address</tt>". Address must be derived from a <a
3548 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3549
3550<h5>Arguments:</h5>
3551
3552<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3553 rest of the arguments indicate the full set of possible destinations that the
3554 address may point to. Blocks are allowed to occur multiple times in the
3555 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3556
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3557<p>This destination list is required so that dataflow analysis has an accurate
3558 understanding of the CFG.</p>
3559
3560<h5>Semantics:</h5>
3561
3562<p>Control transfers to the block specified in the address argument. All
3563 possible destination blocks must be listed in the label list, otherwise this
3564 instruction has undefined behavior. This implies that jumps to labels
3565 defined in other functions have undefined behavior as well.</p>
3566
3567<h5>Implementation:</h5>
3568
3569<p>This is typically implemented with a jump through a register.</p>
3570
3571<h5>Example:</h5>
3572<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3573 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3574</pre>
3575
3576</div>
3577
3578
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3579<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3580<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3581 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3582</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3583
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3584<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3585
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3586<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3587<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3588 &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]3589 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3590</pre>
3591
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3592<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3593<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3594 function, with the possibility of control flow transfer to either the
3595 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3596 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3597 control flow will return to the "normal" label. If the callee (or any
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]3598 indirect callees) returns via the "<a href="#i_resume"><tt>resume</tt></a>"
3599 instruction or other exception handling mechanism, control is interrupted and
3600 continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3601
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3602<p>The '<tt>exception</tt>' label is a
3603 <i><a href="ExceptionHandling.html#overview">landing pad</a></i> for the
3604 exception. As such, '<tt>exception</tt>' label is required to have the
3605 "<a href="#i_landingpad"><tt>landingpad</tt></a>" instruction, which contains
Chad Rosier85f5a1a2011-12-09 02:00:44 +0000[diff] [blame]3606 the information about the behavior of the program after unwinding
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3607 happens, as its first non-PHI instruction. The restrictions on the
3608 "<tt>landingpad</tt>" instruction's tightly couples it to the
3609 "<tt>invoke</tt>" instruction, so that the important information contained
3610 within the "<tt>landingpad</tt>" instruction can't be lost through normal
3611 code motion.</p>
3612
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3613<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3614<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3615
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3616<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3617 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3618 convention</a> the call should use. If none is specified, the call
3619 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3620
3621 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3622 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3623 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3624
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3625 <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]3626 function value being invoked. In most cases, this is a direct function
3627 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3628 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3629
3630 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3631 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3632
3633 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]3634 signature argument types and parameter attributes. All arguments must be
3635 of <a href="#t_firstclass">first class</a> type. If the function
3636 signature indicates the function accepts a variable number of arguments,
3637 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3638
3639 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3640 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3641
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]3642 <li>'<tt>exception label</tt>': the label reached when a callee returns via
3643 the <a href="#i_resume"><tt>resume</tt></a> instruction or other exception
3644 handling mechanism.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3645
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3646 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3647 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3648 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3649</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3650
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3651<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3652<p>This instruction is designed to operate as a standard
3653 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3654 primary difference is that it establishes an association with a label, which
3655 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3656
3657<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3658 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3659 exception. Additionally, this is important for implementation of
3660 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3661
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3662<p>For the purposes of the SSA form, the definition of the value returned by the
3663 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3664 block to the "normal" label. If the callee unwinds then no return value is
3665 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +0000[diff] [blame]3666
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3667<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3668<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3669 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3670 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3671 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3672 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3673</pre>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3674
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3675</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3676
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3677 <!-- _______________________________________________________________________ -->
3678
3679<h4>
3680 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3681</h4>
3682
3683<div>
3684
3685<h5>Syntax:</h5>
3686<pre>
3687 resume &lt;type&gt; &lt;value&gt;
3688</pre>
3689
3690<h5>Overview:</h5>
3691<p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3692 successors.</p>
3693
3694<h5>Arguments:</h5>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3695<p>The '<tt>resume</tt>' instruction requires one argument, which must have the
Bill Wendlinge4ad50b2011-08-03 18:37:32 +0000[diff] [blame]3696 same type as the result of any '<tt>landingpad</tt>' instruction in the same
3697 function.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3698
3699<h5>Semantics:</h5>
3700<p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3701 (in-flight) exception whose unwinding was interrupted with
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3702 a <a href="#i_landingpad"><tt>landingpad</tt></a> instruction.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3703
3704<h5>Example:</h5>
3705<pre>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3706 resume { i8*, i32 } %exn
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3707</pre>
3708
3709</div>
3710
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3711<!-- _______________________________________________________________________ -->
3712
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3713<h4>
3714 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3715</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3716
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3717<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3718
3719<h5>Syntax:</h5>
3720<pre>
3721 unreachable
3722</pre>
3723
3724<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3725<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3726 instruction is used to inform the optimizer that a particular portion of the
3727 code is not reachable. This can be used to indicate that the code after a
3728 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3729
3730<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3731<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3732
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3733</div>
3734
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3735</div>
3736
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3737<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3738<h3>
3739 <a name="binaryops">Binary Operations</a>
3740</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3741
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3742<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3743
3744<p>Binary operators are used to do most of the computation in a program. They
3745 require two operands of the same type, execute an operation on them, and
3746 produce a single value. The operands might represent multiple data, as is
3747 the case with the <a href="#t_vector">vector</a> data type. The result value
3748 has the same type as its operands.</p>
3749
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3750<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3751
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3752<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3753<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3754 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3755</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3756
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3757<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3758
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3759<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3760<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3761 &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]3762 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3763 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3764 &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]3765</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3766
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3767<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3768<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3769
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3770<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3771<p>The two arguments to the '<tt>add</tt>' instruction must
3772 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3773 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3774
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3775<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3776<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3777
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3778<p>If the sum has unsigned overflow, the result returned is the mathematical
3779 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]3780
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3781<p>Because LLVM integers use a two's complement representation, this instruction
3782 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3783
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3784<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3785 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3786 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3787 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3788 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3789
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3790<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3791<pre>
3792 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3793</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3794
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3795</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3796
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3797<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3798<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3799 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3800</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3801
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3802<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3803
3804<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3805<pre>
3806 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3807</pre>
3808
3809<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3810<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3811
3812<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3813<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3814 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3815 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3816
3817<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3818<p>The value produced is the floating point sum of the two operands.</p>
3819
3820<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3821<pre>
3822 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3823</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3824
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3825</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3826
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3827<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3828<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3829 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3830</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3831
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3832<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3833
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3834<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3835<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3836 &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]3837 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3838 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3839 &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]3840</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3841
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3842<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3843<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3844 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3845
3846<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3847 '<tt>neg</tt>' instruction present in most other intermediate
3848 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3849
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3850<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3851<p>The two arguments to the '<tt>sub</tt>' instruction must
3852 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3853 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3854
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3855<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3856<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3857
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3858<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3859 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3860 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3861
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3862<p>Because LLVM integers use a two's complement representation, this instruction
3863 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3864
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3865<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3866 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3867 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3868 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3869 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3870
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3871<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]3872<pre>
3873 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3874 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3875</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3876
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3877</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3878
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3879<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3880<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3881 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3882</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3883
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3884<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3885
3886<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3887<pre>
3888 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3889</pre>
3890
3891<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3892<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3893 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3894
3895<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3896 '<tt>fneg</tt>' instruction present in most other intermediate
3897 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3898
3899<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]3900<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3901 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3902 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3903
3904<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3905<p>The value produced is the floating point difference of the two operands.</p>
3906
3907<h5>Example:</h5>
3908<pre>
3909 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3910 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3911</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3912
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3913</div>
3914
3915<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3916<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3917 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3918</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3919
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3920<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3921
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3922<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3923<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3924 &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]3925 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3926 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3927 &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]3928</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3929
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3930<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3931<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3932
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3933<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3934<p>The two arguments to the '<tt>mul</tt>' instruction must
3935 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3936 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3937
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3938<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3939<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3940
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3941<p>If the result of the multiplication has unsigned overflow, the result
3942 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3943 width of the result.</p>
3944
3945<p>Because LLVM integers use a two's complement representation, and the result
3946 is the same width as the operands, this instruction returns the correct
3947 result for both signed and unsigned integers. If a full product
3948 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3949 be sign-extended or zero-extended as appropriate to the width of the full
3950 product.</p>
3951
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3952<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3953 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3954 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3955 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3956 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3957
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3958<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3959<pre>
3960 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3961</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3962
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3963</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3964
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3965<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3966<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3967 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3968</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3969
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3970<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3971
3972<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3973<pre>
3974 &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]3975</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3976
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3977<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3978<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3979
3980<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3981<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3982 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3983 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3984
3985<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3986<p>The value produced is the floating point product of the two operands.</p>
3987
3988<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3989<pre>
3990 &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]3991</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3992
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3993</div>
3994
3995<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3996<h4>
3997 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3998</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3999
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4000<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4001
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4002<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4003<pre>
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]4004 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4005 &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]4006</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4007
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4008<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4009<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4010
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4011<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4012<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4013 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4014 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4015
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4016<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4017<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4018
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4019<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4020 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
4021
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4022<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4023
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]4024<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4025 <tt>udiv</tt> is a <a href="#poisonvalues">poison value</a> if %op1 is not a
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]4026 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
4027
4028
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4029<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4030<pre>
4031 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4032</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4033
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4034</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4035
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4036<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4037<h4>
4038 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
4039</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4040
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4041<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4042
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4043<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4044<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4045 &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]4046 &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]4047</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4048
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4049<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4050<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4051
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4052<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4053<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4054 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4055 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4056
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4057<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4058<p>The value produced is the signed integer quotient of the two operands rounded
4059 towards zero.</p>
4060
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4061<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4062 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
4063
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4064<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4065 undefined behavior; this is a rare case, but can occur, for example, by doing
4066 a 32-bit division of -2147483648 by -1.</p>
4067
Dan Gohman9c5beed2009-07-22 00:04:19 +0000[diff] [blame]4068<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4069 <tt>sdiv</tt> is a <a href="#poisonvalues">poison value</a> if the result would
Dan Gohman38da9272010-07-11 00:08:34 +0000[diff] [blame]4070 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4071
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4072<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4073<pre>
4074 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4075</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4076
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4077</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4078
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4079<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4080<h4>
4081 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
4082</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4083
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4084<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4085
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4086<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4087<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4088 &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]4089</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4090
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4091<h5>Overview:</h5>
4092<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4093
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4094<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4095<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4096 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4097 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4098
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4099<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4100<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4101
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4102<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4103<pre>
4104 &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]4105</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4106
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4107</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4108
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4109<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4110<h4>
4111 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
4112</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4113
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4114<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4115
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4116<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4117<pre>
4118 &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]4119</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4120
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4121<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4122<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
4123 division of its two arguments.</p>
4124
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4125<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4126<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4127 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4128 values. Both arguments must have identical types.</p>
4129
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4130<h5>Semantics:</h5>
4131<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4132 This instruction always performs an unsigned division to get the
4133 remainder.</p>
4134
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4135<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4136 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
4137
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4138<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4139
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4140<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4141<pre>
4142 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4143</pre>
4144
4145</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4146
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4147<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4148<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4149 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4150</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4151
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4152<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4153
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4154<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4155<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4156 &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]4157</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4158
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4159<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4160<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
4161 division of its two operands. This instruction can also take
4162 <a href="#t_vector">vector</a> versions of the values in which case the
4163 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]4164
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4165<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4166<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4167 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4168 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4169
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4170<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4171<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +0000[diff] [blame]4172 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
4173 <i>modulo</i> operator (where the result is either zero or has the same sign
4174 as the divisor, <tt>op2</tt>) of a value.
4175 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4176 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
4177 Math Forum</a>. For a table of how this is implemented in various languages,
4178 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
4179 Wikipedia: modulo operation</a>.</p>
4180
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4181<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4182 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
4183
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4184<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4185 Overflow also leads to undefined behavior; this is a rare case, but can
4186 occur, for example, by taking the remainder of a 32-bit division of
4187 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
4188 lets srem be implemented using instructions that return both the result of
4189 the division and the remainder.)</p>
4190
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4191<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4192<pre>
4193 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4194</pre>
4195
4196</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4197
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4198<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4199<h4>
4200 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
4201</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4202
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4203<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4204
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4205<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4206<pre>
4207 &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]4208</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4209
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4210<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4211<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
4212 its two operands.</p>
4213
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4214<h5>Arguments:</h5>
4215<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4216 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4217 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4218
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4219<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4220<p>This instruction returns the <i>remainder</i> of a division. The remainder
4221 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4222
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4223<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4224<pre>
4225 &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]4226</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4227
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4228</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]4229
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4230</div>
4231
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4232<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4233<h3>
4234 <a name="bitwiseops">Bitwise Binary Operations</a>
4235</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4236
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4237<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4238
4239<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
4240 program. They are generally very efficient instructions and can commonly be
4241 strength reduced from other instructions. They require two operands of the
4242 same type, execute an operation on them, and produce a single value. The
4243 resulting value is the same type as its operands.</p>
4244
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4245<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4246<h4>
4247 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
4248</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4249
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4250<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4251
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4252<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4253<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4254 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4255 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4256 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4257 &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]4258</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4259
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4260<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4261<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
4262 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4263
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4264<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4265<p>Both arguments to the '<tt>shl</tt>' instruction must be the
4266 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
4267 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4268
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4269<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4270<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
4271 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
4272 is (statically or dynamically) negative or equal to or larger than the number
4273 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4274 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4275 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4276
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4277<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4278 <a href="#poisonvalues">poison value</a> if it shifts out any non-zero bits. If
Chris Lattner66298c12011-02-09 16:44:44 +0000[diff] [blame]4279 the <tt>nsw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4280 <a href="#poisonvalues">poison value</a> if it shifts out any bits that disagree
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4281 with the resultant sign bit. As such, NUW/NSW have the same semantics as
4282 they would if the shift were expressed as a mul instruction with the same
4283 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
4284
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4285<h5>Example:</h5>
4286<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4287 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
4288 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
4289 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4290 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4291 &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]4292</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4293
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4294</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4295
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4296<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4297<h4>
4298 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
4299</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4300
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4301<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4302
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4303<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4304<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4305 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4306 &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]4307</pre>
4308
4309<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4310<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
4311 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4312
4313<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4314<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4315 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4316 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4317
4318<h5>Semantics:</h5>
4319<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4320 significant bits of the result will be filled with zero bits after the shift.
4321 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4322 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4323 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4324 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4325
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4326<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4327 <tt>lshr</tt> is a <a href="#poisonvalues">poison value</a> if any of the bits
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4328 shifted out are non-zero.</p>
4329
4330
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4331<h5>Example:</h5>
4332<pre>
4333 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4334 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4335 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4336 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4337 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4338 &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]4339</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4340
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4341</div>
4342
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4343<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4344<h4>
4345 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4346</h4>
4347
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4348<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4349
4350<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4351<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4352 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4353 &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]4354</pre>
4355
4356<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4357<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4358 operand shifted to the right a specified number of bits with sign
4359 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4360
4361<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4362<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4363 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4364 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4365
4366<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4367<p>This instruction always performs an arithmetic shift right operation, The
4368 most significant bits of the result will be filled with the sign bit
4369 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4370 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4371 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4372 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4373
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4374<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4375 <tt>ashr</tt> is a <a href="#poisonvalues">poison value</a> if any of the bits
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4376 shifted out are non-zero.</p>
4377
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4378<h5>Example:</h5>
4379<pre>
4380 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4381 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4382 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4383 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4384 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4385 &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]4386</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4387
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4388</div>
4389
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4390<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4391<h4>
4392 <a name="i_and">'<tt>and</tt>' Instruction</a>
4393</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4394
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4395<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4396
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4397<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4398<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4399 &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]4400</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4401
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4402<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4403<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4404 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4405
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4406<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4407<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4408 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4409 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4410
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4411<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4412<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4413
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4414<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4415 <tbody>
4416 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4417 <th>In0</th>
4418 <th>In1</th>
4419 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4420 </tr>
4421 <tr>
4422 <td>0</td>
4423 <td>0</td>
4424 <td>0</td>
4425 </tr>
4426 <tr>
4427 <td>0</td>
4428 <td>1</td>
4429 <td>0</td>
4430 </tr>
4431 <tr>
4432 <td>1</td>
4433 <td>0</td>
4434 <td>0</td>
4435 </tr>
4436 <tr>
4437 <td>1</td>
4438 <td>1</td>
4439 <td>1</td>
4440 </tr>
4441 </tbody>
4442</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4443
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4444<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4445<pre>
4446 &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]4447 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4448 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4449</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4450</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4451<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4452<h4>
4453 <a name="i_or">'<tt>or</tt>' Instruction</a>
4454</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4455
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4456<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4457
4458<h5>Syntax:</h5>
4459<pre>
4460 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4461</pre>
4462
4463<h5>Overview:</h5>
4464<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4465 two operands.</p>
4466
4467<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4468<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4469 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4470 values. Both arguments must have identical types.</p>
4471
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4472<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4473<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4474
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4475<table border="1" cellspacing="0" cellpadding="4">
4476 <tbody>
4477 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4478 <th>In0</th>
4479 <th>In1</th>
4480 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4481 </tr>
4482 <tr>
4483 <td>0</td>
4484 <td>0</td>
4485 <td>0</td>
4486 </tr>
4487 <tr>
4488 <td>0</td>
4489 <td>1</td>
4490 <td>1</td>
4491 </tr>
4492 <tr>
4493 <td>1</td>
4494 <td>0</td>
4495 <td>1</td>
4496 </tr>
4497 <tr>
4498 <td>1</td>
4499 <td>1</td>
4500 <td>1</td>
4501 </tr>
4502 </tbody>
4503</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4504
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4505<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4506<pre>
4507 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4508 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4509 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4510</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4511
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4512</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4513
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4514<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4515<h4>
4516 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4517</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4518
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4519<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4520
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4521<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4522<pre>
4523 &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]4524</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4525
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4526<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4527<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4528 its two operands. The <tt>xor</tt> is used to implement the "one's
4529 complement" operation, which is the "~" operator in C.</p>
4530
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4531<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4532<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4533 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4534 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4535
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4536<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4537<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4538
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4539<table border="1" cellspacing="0" cellpadding="4">
4540 <tbody>
4541 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4542 <th>In0</th>
4543 <th>In1</th>
4544 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4545 </tr>
4546 <tr>
4547 <td>0</td>
4548 <td>0</td>
4549 <td>0</td>
4550 </tr>
4551 <tr>
4552 <td>0</td>
4553 <td>1</td>
4554 <td>1</td>
4555 </tr>
4556 <tr>
4557 <td>1</td>
4558 <td>0</td>
4559 <td>1</td>
4560 </tr>
4561 <tr>
4562 <td>1</td>
4563 <td>1</td>
4564 <td>0</td>
4565 </tr>
4566 </tbody>
4567</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4568
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4569<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4570<pre>
4571 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4572 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4573 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4574 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4575</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4576
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4577</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4578
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4579</div>
4580
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4581<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4582<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4583 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4584</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4585
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4586<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4587
4588<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4589 target-independent manner. These instructions cover the element-access and
4590 vector-specific operations needed to process vectors effectively. While LLVM
4591 does directly support these vector operations, many sophisticated algorithms
4592 will want to use target-specific intrinsics to take full advantage of a
4593 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4594
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4595<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4596<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4597 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4598</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4599
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4600<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4601
4602<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4603<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4604 &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]4605</pre>
4606
4607<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4608<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4609 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4610
4611
4612<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4613<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4614 of <a href="#t_vector">vector</a> type. The second operand is an index
4615 indicating the position from which to extract the element. The index may be
4616 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4617
4618<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4619<p>The result is a scalar of the same type as the element type of
4620 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4621 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4622 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4623
4624<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4625<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4626 &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]4627</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4628
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4629</div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4630
4631<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4632<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4633 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4634</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4635
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4636<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4637
4638<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4639<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +0000[diff] [blame]4640 &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]4641</pre>
4642
4643<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4644<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4645 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4646
4647<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4648<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4649 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4650 whose type must equal the element type of the first operand. The third
4651 operand is an index indicating the position at which to insert the value.
4652 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4653
4654<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4655<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4656 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4657 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4658 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4659
4660<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4661<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4662 &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]4663</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4664
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4665</div>
4666
4667<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4668<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4669 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4670</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4671
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4672<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4673
4674<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4675<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4676 &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]4677</pre>
4678
4679<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4680<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4681 from two input vectors, returning a vector with the same element type as the
4682 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4683
4684<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4685<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4686 with types that match each other. The third argument is a shuffle mask whose
4687 element type is always 'i32'. The result of the instruction is a vector
4688 whose length is the same as the shuffle mask and whose element type is the
4689 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4690
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4691<p>The shuffle mask operand is required to be a constant vector with either
4692 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4693
4694<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4695<p>The elements of the two input vectors are numbered from left to right across
4696 both of the vectors. The shuffle mask operand specifies, for each element of
4697 the result vector, which element of the two input vectors the result element
4698 gets. The element selector may be undef (meaning "don't care") and the
4699 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4700
4701<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4702<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4703 &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]4704 &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]4705 &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]4706 &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]4707 &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]4708 &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]4709 &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]4710 &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]4711</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4712
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4713</div>
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]4714
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4715</div>
4716
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4717<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4718<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4719 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4720</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4721
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4722<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4723
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4724<p>LLVM supports several instructions for working with
4725 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4726
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4727<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4728<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4729 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4730</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4731
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4732<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4733
4734<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4735<pre>
4736 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4737</pre>
4738
4739<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4740<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4741 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4742
4743<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4744<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4745 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4746 <a href="#t_array">array</a> type. The operands are constant indices to
4747 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4748 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4749 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4750 <ul>
4751 <li>Since the value being indexed is not a pointer, the first index is
4752 omitted and assumed to be zero.</li>
4753 <li>At least one index must be specified.</li>
4754 <li>Not only struct indices but also array indices must be in
4755 bounds.</li>
4756 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4757
4758<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4759<p>The result is the value at the position in the aggregate specified by the
4760 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4761
4762<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4763<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4764 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4765</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4766
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4767</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4768
4769<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4770<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4771 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4772</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4773
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4774<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4775
4776<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4777<pre>
Bill Wendling194229e2011-07-26 20:42:28 +0000[diff] [blame]4778 &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]4779</pre>
4780
4781<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4782<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4783 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4784
4785<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4786<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4787 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4788 <a href="#t_array">array</a> type. The second operand is a first-class
4789 value to insert. The following operands are constant indices indicating
4790 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]4791 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4792 value to insert must have the same type as the value identified by the
4793 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4794
4795<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4796<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4797 that of <tt>val</tt> except that the value at the position specified by the
4798 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4799
4800<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4801<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +0000[diff] [blame]4802 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4803 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4804 %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]4805</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4806
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4807</div>
4808
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4809</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4810
4811<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4812<h3>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4813 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4814</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4815
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4816<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4817
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4818<p>A key design point of an SSA-based representation is how it represents
4819 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]4820 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4821 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4822
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4823<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4824<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4825 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4826</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4827
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4828<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4829
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4830<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4831<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +0000[diff] [blame]4832 &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]4833</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4834
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4835<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4836<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4837 currently executing function, to be automatically released when this function
4838 returns to its caller. The object is always allocated in the generic address
4839 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4840
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4841<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4842<p>The '<tt>alloca</tt>' instruction
4843 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4844 runtime stack, returning a pointer of the appropriate type to the program.
4845 If "NumElements" is specified, it is the number of elements allocated,
4846 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4847 specified, the value result of the allocation is guaranteed to be aligned to
4848 at least that boundary. If not specified, or if zero, the target can choose
4849 to align the allocation on any convenient boundary compatible with the
4850 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4851
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4852<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4853
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4854<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +0000[diff] [blame]4855<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4856 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4857 memory is automatically released when the function returns. The
4858 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4859 variables that must have an address available. When the function returns
4860 (either with the <tt><a href="#i_ret">ret</a></tt>
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]4861 or <tt><a href="#i_resume">resume</a></tt> instructions), the memory is
Nick Lewycky84a1d232012-02-29 08:26:44 +0000[diff] [blame]4862 reclaimed. Allocating zero bytes is legal, but the result is undefined.
4863 The order in which memory is allocated (ie., which way the stack grows) is
4864 not specified, and relational comparisons involving '<tt>alloca</tt>'s are
4865 undefined.</p>
4866
4867<p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4868
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4869<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4870<pre>
Dan Gohman81e21672009-01-04 23:49:44 +0000[diff] [blame]4871 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4872 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4873 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4874 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4875</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4876
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4877</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4878
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4879<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4880<h4>
4881 <a name="i_load">'<tt>load</tt>' Instruction</a>
4882</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4883
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4884<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4885
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4886<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4887<pre>
Pete Cooperf95acc62012-02-10 18:13:54 +0000[diff] [blame]4888 &lt;result&gt; = load [volatile] &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;][, !invariant.load !&lt;index&gt;]
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]4889 &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]4890 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4891</pre>
4892
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4893<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4894<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4895
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4896<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4897<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4898 from which to load. The pointer must point to
4899 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4900 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4901 number or order of execution of this <tt>load</tt> with other <a
4902 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4903
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]4904<p>If the <code>load</code> is marked as <code>atomic</code>, it takes an extra
4905 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4906 argument. The <code>release</code> and <code>acq_rel</code> orderings are
4907 not valid on <code>load</code> instructions. Atomic loads produce <a
4908 href="#memorymodel">defined</a> results when they may see multiple atomic
4909 stores. The type of the pointee must be an integer type whose bit width
4910 is a power of two greater than or equal to eight and less than or equal
4911 to a target-specific size limit. <code>align</code> must be explicitly
4912 specified on atomic loads, and the load has undefined behavior if the
4913 alignment is not set to a value which is at least the size in bytes of
4914 the pointee. <code>!nontemporal</code> does not have any defined semantics
4915 for atomic loads.</p>
4916
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4917<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4918 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]4919 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4920 alignment for the target. It is the responsibility of the code emitter to
4921 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4922 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4923 produce less efficient code. An alignment of 1 is always safe.</p>
4924
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4925<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4926 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4927 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4928 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4929 and code generator that this load is not expected to be reused in the cache.
4930 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4931 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4932
Pete Cooperf95acc62012-02-10 18:13:54 +0000[diff] [blame]4933<p>The optional <tt>!invariant.load</tt> metadata must reference a single
4934 metatadata name &lt;index&gt; corresponding to a metadata node with no
4935 entries. The existence of the <tt>!invariant.load</tt> metatadata on the
4936 instruction tells the optimizer and code generator that this load address
4937 points to memory which does not change value during program execution.
4938 The optimizer may then move this load around, for example, by hoisting it
4939 out of loops using loop invariant code motion.</p>
4940
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4941<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4942<p>The location of memory pointed to is loaded. If the value being loaded is of
4943 scalar type then the number of bytes read does not exceed the minimum number
4944 of bytes needed to hold all bits of the type. For example, loading an
4945 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4946 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4947 is undefined if the value was not originally written using a store of the
4948 same type.</p>
4949
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4950<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4951<pre>
4952 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4953 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4954 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4955</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4956
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4957</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4958
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4959<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4960<h4>
4961 <a name="i_store">'<tt>store</tt>' Instruction</a>
4962</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4963
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4964<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4965
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4966<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4967<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4968 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>
4969 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]4970</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4971
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4972<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4973<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4974
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4975<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4976<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4977 and an address at which to store it. The type of the
4978 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4979 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4980 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4981 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4982 order of execution of this <tt>store</tt> with other <a
4983 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4984
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]4985<p>If the <code>store</code> is marked as <code>atomic</code>, it takes an extra
4986 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4987 argument. The <code>acquire</code> and <code>acq_rel</code> orderings aren't
4988 valid on <code>store</code> instructions. Atomic loads produce <a
4989 href="#memorymodel">defined</a> results when they may see multiple atomic
4990 stores. The type of the pointee must be an integer type whose bit width
4991 is a power of two greater than or equal to eight and less than or equal
4992 to a target-specific size limit. <code>align</code> must be explicitly
4993 specified on atomic stores, and the store has undefined behavior if the
4994 alignment is not set to a value which is at least the size in bytes of
4995 the pointee. <code>!nontemporal</code> does not have any defined semantics
4996 for atomic stores.</p>
4997
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4998<p>The optional constant "align" argument specifies the alignment of the
4999 operation (that is, the alignment of the memory address). A value of 0 or an
5000 omitted "align" argument means that the operation has the preferential
5001 alignment for the target. It is the responsibility of the code emitter to
5002 ensure that the alignment information is correct. Overestimating the
5003 alignment results in an undefined behavior. Underestimating the alignment may
5004 produce less efficient code. An alignment of 1 is always safe.</p>
5005
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5006<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5007 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5008 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5009 instruction tells the optimizer and code generator that this load is
5010 not expected to be reused in the cache. The code generator may
5011 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5012 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5013
5014
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]5015<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5016<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
5017 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
5018 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
5019 does not exceed the minimum number of bytes needed to hold all bits of the
5020 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
5021 writing a value of a type like <tt>i20</tt> with a size that is not an
5022 integral number of bytes, it is unspecified what happens to the extra bits
5023 that do not belong to the type, but they will typically be overwritten.</p>
5024
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5025<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5026<pre>
5027 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]5028 store i32 3, i32* %ptr <i>; yields {void}</i>
5029 %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]5030</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5031
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5032</div>
5033
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5034<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5035<h4>
5036<a name="i_fence">'<tt>fence</tt>' Instruction</a>
5037</h4>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5038
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5039<div>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5040
5041<h5>Syntax:</h5>
5042<pre>
5043 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
5044</pre>
5045
5046<h5>Overview:</h5>
5047<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
5048between operations.</p>
5049
5050<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
5051href="#ordering">ordering</a> argument which defines what
5052<i>synchronizes-with</i> edges they add. They can only be given
5053<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
5054<code>seq_cst</code> orderings.</p>
5055
5056<h5>Semantics:</h5>
5057<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
5058semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
5059<code>acquire</code> ordering semantics if and only if there exist atomic
5060operations <var>X</var> and <var>Y</var>, both operating on some atomic object
5061<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
5062<var>X</var> modifies <var>M</var> (either directly or through some side effect
5063of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
5064<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
5065<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
5066than an explicit <code>fence</code>, one (but not both) of the atomic operations
5067<var>X</var> or <var>Y</var> might provide a <code>release</code> or
5068<code>acquire</code> (resp.) ordering constraint and still
5069<i>synchronize-with</i> the explicit <code>fence</code> and establish the
5070<i>happens-before</i> edge.</p>
5071
5072<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
5073having both <code>acquire</code> and <code>release</code> semantics specified
5074above, participates in the global program order of other <code>seq_cst</code>
5075operations and/or fences.</p>
5076
5077<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
5078specifies that the fence only synchronizes with other fences in the same
5079thread. (This is useful for interacting with signal handlers.)</p>
5080
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5081<h5>Example:</h5>
5082<pre>
5083 fence acquire <i>; yields {void}</i>
5084 fence singlethread seq_cst <i>; yields {void}</i>
5085</pre>
5086
5087</div>
5088
5089<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5090<h4>
5091<a name="i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a>
5092</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5093
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5094<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5095
5096<h5>Syntax:</h5>
5097<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5098 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]5099</pre>
5100
5101<h5>Overview:</h5>
5102<p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
5103It loads a value in memory and compares it to a given value. If they are
5104equal, it stores a new value into the memory.</p>
5105
5106<h5>Arguments:</h5>
5107<p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
5108address to operate on, a value to compare to the value currently be at that
5109address, and a new value to place at that address if the compared values are
5110equal. The type of '<var>&lt;cmp&gt;</var>' must be an integer type whose
5111bit width is a power of two greater than or equal to eight and less than
5112or equal to a target-specific size limit. '<var>&lt;cmp&gt;</var>' and
5113'<var>&lt;new&gt;</var>' must have the same type, and the type of
5114'<var>&lt;pointer&gt;</var>' must be a pointer to that type. If the
5115<code>cmpxchg</code> is marked as <code>volatile</code>, then the
5116optimizer is not allowed to modify the number or order of execution
5117of this <code>cmpxchg</code> with other <a href="#volatile">volatile
5118operations</a>.</p>
5119
5120<!-- FIXME: Extend allowed types. -->
5121
5122<p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
5123<code>cmpxchg</code> synchronizes with other atomic operations.</p>
5124
5125<p>The optional "<code>singlethread</code>" argument declares that the
5126<code>cmpxchg</code> is only atomic with respect to code (usually signal
5127handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
5128cmpxchg is atomic with respect to all other code in the system.</p>
5129
5130<p>The pointer passed into cmpxchg must have alignment greater than or equal to
5131the size in memory of the operand.
5132
5133<h5>Semantics:</h5>
5134<p>The contents of memory at the location specified by the
5135'<tt>&lt;pointer&gt;</tt>' operand is read and compared to
5136'<tt>&lt;cmp&gt;</tt>'; if the read value is the equal,
5137'<tt>&lt;new&gt;</tt>' is written. The original value at the location
5138is returned.
5139
5140<p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
5141purpose of identifying <a href="#release_sequence">release sequences</a>. A
5142failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
5143parameter determined by dropping any <code>release</code> part of the
5144<code>cmpxchg</code>'s ordering.</p>
5145
5146<!--
5147FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
5148optimization work on ARM.)
5149
5150FIXME: Is a weaker ordering constraint on failure helpful in practice?
5151-->
5152
5153<h5>Example:</h5>
5154<pre>
5155entry:
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5156 %orig = atomic <a href="#i_load">load</a> i32* %ptr unordered <i>; yields {i32}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5157 <a href="#i_br">br</a> label %loop
5158
5159loop:
5160 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
5161 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5162 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5163 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
5164 <a href="#i_br">br</a> i1 %success, label %done, label %loop
5165
5166done:
5167 ...
5168</pre>
5169
5170</div>
5171
5172<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5173<h4>
5174<a name="i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a>
5175</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5176
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5177<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5178
5179<h5>Syntax:</h5>
5180<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]5181 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]5182</pre>
5183
5184<h5>Overview:</h5>
5185<p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
5186
5187<h5>Arguments:</h5>
5188<p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
5189operation to apply, an address whose value to modify, an argument to the
5190operation. The operation must be one of the following keywords:</p>
5191<ul>
5192 <li>xchg</li>
5193 <li>add</li>
5194 <li>sub</li>
5195 <li>and</li>
5196 <li>nand</li>
5197 <li>or</li>
5198 <li>xor</li>
5199 <li>max</li>
5200 <li>min</li>
5201 <li>umax</li>
5202 <li>umin</li>
5203</ul>
5204
5205<p>The type of '<var>&lt;value&gt;</var>' must be an integer type whose
5206bit width is a power of two greater than or equal to eight and less than
5207or equal to a target-specific size limit. The type of the
5208'<code>&lt;pointer&gt;</code>' operand must be a pointer to that type.
5209If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
5210optimizer is not allowed to modify the number or order of execution of this
5211<code>atomicrmw</code> with other <a href="#volatile">volatile
5212 operations</a>.</p>
5213
5214<!-- FIXME: Extend allowed types. -->
5215
5216<h5>Semantics:</h5>
5217<p>The contents of memory at the location specified by the
5218'<tt>&lt;pointer&gt;</tt>' operand are atomically read, modified, and written
5219back. The original value at the location is returned. The modification is
5220specified by the <var>operation</var> argument:</p>
5221
5222<ul>
5223 <li>xchg: <code>*ptr = val</code></li>
5224 <li>add: <code>*ptr = *ptr + val</code></li>
5225 <li>sub: <code>*ptr = *ptr - val</code></li>
5226 <li>and: <code>*ptr = *ptr &amp; val</code></li>
5227 <li>nand: <code>*ptr = ~(*ptr &amp; val)</code></li>
5228 <li>or: <code>*ptr = *ptr | val</code></li>
5229 <li>xor: <code>*ptr = *ptr ^ val</code></li>
5230 <li>max: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using a signed comparison)</li>
5231 <li>min: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using a signed comparison)</li>
5232 <li>umax: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5233 <li>umin: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5234</ul>
5235
5236<h5>Example:</h5>
5237<pre>
5238 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
5239</pre>
5240
5241</div>
5242
5243<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5244<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5245 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5246</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5247
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5248<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5249
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5250<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5251<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5252 &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]5253 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5254 &lt;result&gt; = getelementptr &lt;ptr vector&gt; ptrval, &lt;vector index type&gt; idx
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5255</pre>
5256
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5257<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5258<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5259 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
5260 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5261
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5262<h5>Arguments:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5263<p>The first argument is always a pointer or a vector of pointers,
5264 and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]5265 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5266 elements of the aggregate object are indexed. The interpretation of each
5267 index is dependent on the type being indexed into. The first index always
5268 indexes the pointer value given as the first argument, the second index
5269 indexes a value of the type pointed to (not necessarily the value directly
5270 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5271 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]5272 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5273 can never be pointers, since that would require loading the pointer before
5274 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5275
5276<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]5277 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5278 integer <b>constants</b> are allowed. When indexing into an array, pointer
5279 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]5280 constant. These integers are treated as signed values where relevant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5281
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5282<p>For example, let's consider a C code fragment and how it gets compiled to
5283 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5284
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5285<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5286struct RT {
5287 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5288 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5289 char C;
5290};
5291struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5292 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5293 double Y;
5294 struct RT Z;
5295};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5296
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5297int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5298 return &amp;s[1].Z.B[5][13];
5299}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5300</pre>
5301
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5302<p>The LLVM code generated by Clang is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5303
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5304<pre class="doc_code">
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5305%struct.RT = <a href="#namedtypes">type</a> { i8, [10 x [20 x i32]], i8 }
5306%struct.ST = <a href="#namedtypes">type</a> { i32, double, %struct.RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5307
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5308define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5309entry:
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5310 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5311 ret i32* %arrayidx
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5312}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5313</pre>
5314
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5315<h5>Semantics:</h5>
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5316<p>In the example above, the first index is indexing into the
5317 '<tt>%struct.ST*</tt>' type, which is a pointer, yielding a
5318 '<tt>%struct.ST</tt>' = '<tt>{ i32, double, %struct.RT }</tt>' type, a
5319 structure. The second index indexes into the third element of the structure,
5320 yielding a '<tt>%struct.RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]], i8 }</tt>'
5321 type, another structure. The third index indexes into the second element of
5322 the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an array. The
5323 two dimensions of the array are subscripted into, yielding an '<tt>i32</tt>'
5324 type. The '<tt>getelementptr</tt>' instruction returns a pointer to this
5325 element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5326
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5327<p>Note that it is perfectly legal to index partially through a structure,
5328 returning a pointer to an inner element. Because of this, the LLVM code for
5329 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5330
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5331<pre class="doc_code">
5332define i32* @foo(%struct.ST* %s) {
5333 %t1 = getelementptr %struct.ST* %s, i32 1 <i>; yields %struct.ST*:%t1</i>
5334 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 <i>; yields %struct.RT*:%t2</i>
5335 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
5336 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
5337 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
5338 ret i32* %t5
5339}
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5340</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]5341
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5342<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]5343 <tt>getelementptr</tt> is a <a href="#poisonvalues">poison value</a> if the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]5344 base pointer is not an <i>in bounds</i> address of an allocated object,
5345 or if any of the addresses that would be formed by successive addition of
5346 the offsets implied by the indices to the base address with infinitely
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5347 precise signed arithmetic are not an <i>in bounds</i> address of that
5348 allocated object. The <i>in bounds</i> addresses for an allocated object
5349 are all the addresses that point into the object, plus the address one
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5350 byte past the end.
5351 In cases where the base is a vector of pointers the <tt>inbounds</tt> keyword
5352 applies to each of the computations element-wise. </p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5353
5354<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]5355 the base address with silently-wrapping two's complement arithmetic. If the
5356 offsets have a different width from the pointer, they are sign-extended or
5357 truncated to the width of the pointer. The result value of the
5358 <tt>getelementptr</tt> may be outside the object pointed to by the base
5359 pointer. The result value may not necessarily be used to access memory
5360 though, even if it happens to point into allocated storage. See the
5361 <a href="#pointeraliasing">Pointer Aliasing Rules</a> section for more
5362 information.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5363
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5364<p>The getelementptr instruction is often confusing. For some more insight into
5365 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]5366
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5367<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5368<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5369 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5370 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5371 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5372 %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]5373 <i>; yields i8*:eptr</i>
5374 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +0000[diff] [blame]5375 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +0000[diff] [blame]5376 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5377</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5378
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5379<p>In cases where the pointer argument is a vector of pointers, only a
5380 single index may be used, and the number of vector elements has to be
5381 the same. For example: </p>
5382<pre class="doc_code">
5383 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5384</pre>
5385
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5386</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5387
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5388</div>
5389
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5390<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5391<h3>
5392 <a name="convertops">Conversion Operations</a>
5393</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5394
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5395<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5396
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5397<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5398 which all take a single operand and a type. They perform various bit
5399 conversions on the operand.</p>
5400
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5401<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5402<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5403 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5404</h4>
5405
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5406<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5407
5408<h5>Syntax:</h5>
5409<pre>
5410 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5411</pre>
5412
5413<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5414<p>The '<tt>trunc</tt>' instruction truncates its operand to the
5415 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5416
5417<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5418<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5419 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5420 of the same number of integers.
5421 The bit size of the <tt>value</tt> must be larger than
5422 the bit size of the destination type, <tt>ty2</tt>.
5423 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5424
5425<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5426<p>The '<tt>trunc</tt>' instruction truncates the high order bits
5427 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5428 source size must be larger than the destination size, <tt>trunc</tt> cannot
5429 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5430
5431<h5>Example:</h5>
5432<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5433 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5434 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5435 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5436 %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]5437</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5438
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5439</div>
5440
5441<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5442<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5443 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5444</h4>
5445
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5446<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5447
5448<h5>Syntax:</h5>
5449<pre>
5450 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5451</pre>
5452
5453<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5454<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5455 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5456
5457
5458<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5459<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5460 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5461 of the same number of integers.
5462 The bit size of the <tt>value</tt> must be smaller than
5463 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5464 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5465
5466<h5>Semantics:</h5>
5467<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]5468 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5469
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]5470<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]5471
5472<h5>Example:</h5>
5473<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5474 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5475 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5476 %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]5477</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5478
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5479</div>
5480
5481<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5482<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5483 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5484</h4>
5485
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5486<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5487
5488<h5>Syntax:</h5>
5489<pre>
5490 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5491</pre>
5492
5493<h5>Overview:</h5>
5494<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5495
5496<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5497<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5498 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5499 of the same number of integers.
5500 The bit size of the <tt>value</tt> must be smaller than
5501 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5502 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5503
5504<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5505<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5506 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5507 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5508
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5509<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]5510
5511<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5512<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5513 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5514 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5515 %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]5516</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5517
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5518</div>
5519
5520<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5521<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5522 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5523</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5524
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5525<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5526
5527<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5528<pre>
5529 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5530</pre>
5531
5532<h5>Overview:</h5>
5533<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5534 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5535
5536<h5>Arguments:</h5>
5537<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5538 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5539 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]5540 <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]5541 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5542
5543<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5544<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5545 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5546 <a href="#t_floating">floating point</a> type. If the value cannot fit
5547 within the destination type, <tt>ty2</tt>, then the results are
5548 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5549
5550<h5>Example:</h5>
5551<pre>
5552 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5553 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5554</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5555
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5556</div>
5557
5558<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5559<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5560 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5561</h4>
5562
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5563<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5564
5565<h5>Syntax:</h5>
5566<pre>
5567 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5568</pre>
5569
5570<h5>Overview:</h5>
5571<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]5572 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5573
5574<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5575<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5576 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5577 a <a href="#t_floating">floating point</a> type to cast it to. The source
5578 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5579
5580<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5581<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5582 <a href="#t_floating">floating point</a> type to a larger
5583 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5584 used to make a <i>no-op cast</i> because it always changes bits. Use
5585 <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]5586
5587<h5>Example:</h5>
5588<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +0000[diff] [blame]5589 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5590 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5591</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5592
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5593</div>
5594
5595<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5596<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]5597 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5598</h4>
5599
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5600<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5601
5602<h5>Syntax:</h5>
5603<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5604 &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]5605</pre>
5606
5607<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5608<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5609 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5610
5611<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5612<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5613 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5614 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5615 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5616 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5617
5618<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5619<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5620 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5621 towards zero) unsigned integer value. If the value cannot fit
5622 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5623
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5624<h5>Example:</h5>
5625<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5626 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5627 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5628 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5629</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5630
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5631</div>
5632
5633<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5634<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5635 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5636</h4>
5637
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5638<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5639
5640<h5>Syntax:</h5>
5641<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5642 &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]5643</pre>
5644
5645<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5646<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5647 <a href="#t_floating">floating point</a> <tt>value</tt> to
5648 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5649
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5650<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5651<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5652 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5653 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5654 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5655 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5656
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5657<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5658<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5659 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5660 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5661 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5662
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5663<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5664<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5665 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5666 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5667 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5668</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5669
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5670</div>
5671
5672<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5673<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5674 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5675</h4>
5676
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5677<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5678
5679<h5>Syntax:</h5>
5680<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5681 &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]5682</pre>
5683
5684<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5685<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5686 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5687
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5688<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5689<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]5690 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5691 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5692 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5693 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5694
5695<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5696<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5697 integer quantity and converts it to the corresponding floating point
5698 value. If the value cannot fit in the floating point value, the results are
5699 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5700
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5701<h5>Example:</h5>
5702<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5703 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5704 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5705</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5706
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5707</div>
5708
5709<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5710<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5711 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5712</h4>
5713
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5714<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5715
5716<h5>Syntax:</h5>
5717<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5718 &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]5719</pre>
5720
5721<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5722<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5723 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5724
5725<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5726<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]5727 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5728 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5729 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5730 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5731
5732<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5733<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5734 quantity and converts it to the corresponding floating point value. If the
5735 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5736
5737<h5>Example:</h5>
5738<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5739 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5740 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5741</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5742
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5743</div>
5744
5745<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5746<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5747 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5748</h4>
5749
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5750<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5751
5752<h5>Syntax:</h5>
5753<pre>
5754 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5755</pre>
5756
5757<h5>Overview:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5758<p>The '<tt>ptrtoint</tt>' instruction converts the pointer or a vector of
5759 pointers <tt>value</tt> to
5760 the integer (or vector of integers) type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5761
5762<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5763<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5764 must be a a value of type <a href="#t_pointer">pointer</a> or a vector of
5765 pointers, and a type to cast it to
5766 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> or a vector
5767 of integers type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5768
5769<h5>Semantics:</h5>
5770<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5771 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5772 truncating or zero extending that value to the size of the integer type. If
5773 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5774 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5775 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5776 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5777
5778<h5>Example:</h5>
5779<pre>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5780 %X = ptrtoint i32* %P to i8 <i>; yields truncation on 32-bit architecture</i>
5781 %Y = ptrtoint i32* %P to i64 <i>; yields zero extension on 32-bit architecture</i>
5782 %Z = ptrtoint &lt;4 x i32*&gt; %P to &lt;4 x i64&gt;<i>; yields vector zero extension for a vector of addresses on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5783</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5784
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5785</div>
5786
5787<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5788<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5789 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5790</h4>
5791
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5792<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5793
5794<h5>Syntax:</h5>
5795<pre>
5796 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5797</pre>
5798
5799<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5800<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5801 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5802
5803<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]5804<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5805 value to cast, and a type to cast it to, which must be a
5806 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5807
5808<h5>Semantics:</h5>
5809<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5810 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5811 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5812 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5813 than the size of a pointer then a zero extension is done. If they are the
5814 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5815
5816<h5>Example:</h5>
5817<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5818 %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]5819 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5820 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5821 %Z = inttoptr &lt;4 x i32&gt; %G to &lt;4 x i8*&gt;<i>; yields truncation of vector G to four pointers</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5822</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5823
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5824</div>
5825
5826<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5827<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5828 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5829</h4>
5830
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5831<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5832
5833<h5>Syntax:</h5>
5834<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5835 &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]5836</pre>
5837
5838<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5839<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5840 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5841
5842<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5843<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5844 non-aggregate first class value, and a type to cast it to, which must also be
5845 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5846 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5847 identical. If the source type is a pointer, the destination type must also be
5848 a pointer. This instruction supports bitwise conversion of vectors to
5849 integers and to vectors of other types (as long as they have the same
5850 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5851
5852<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5853<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5854 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5855 this conversion. The conversion is done as if the <tt>value</tt> had been
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5856 stored to memory and read back as type <tt>ty2</tt>.
5857 Pointer (or vector of pointers) types may only be converted to other pointer
5858 (or vector of pointers) types with this instruction. To convert
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5859 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5860 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5861
5862<h5>Example:</h5>
5863<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5864 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5865 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5866 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
5867 %Z = bitcast &lt;2 x i32*&gt; %V to &lt;2 x i64*&gt; <i>; yields &lt;2 x i64*&gt;</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5868</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5869
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5870</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5871
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5872</div>
5873
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5874<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5875<h3>
5876 <a name="otherops">Other Operations</a>
5877</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5878
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5879<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5880
5881<p>The instructions in this category are the "miscellaneous" instructions, which
5882 defy better classification.</p>
5883
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5884<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5885<h4>
5886 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5887</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5888
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5889<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5890
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5891<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5892<pre>
5893 &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]5894</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5895
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5896<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5897<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5898 boolean values based on comparison of its two integer, integer vector,
5899 pointer, or pointer vector operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5900
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5901<h5>Arguments:</h5>
5902<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5903 the condition code indicating the kind of comparison to perform. It is not a
5904 value, just a keyword. The possible condition code are:</p>
5905
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5906<ol>
5907 <li><tt>eq</tt>: equal</li>
5908 <li><tt>ne</tt>: not equal </li>
5909 <li><tt>ugt</tt>: unsigned greater than</li>
5910 <li><tt>uge</tt>: unsigned greater or equal</li>
5911 <li><tt>ult</tt>: unsigned less than</li>
5912 <li><tt>ule</tt>: unsigned less or equal</li>
5913 <li><tt>sgt</tt>: signed greater than</li>
5914 <li><tt>sge</tt>: signed greater or equal</li>
5915 <li><tt>slt</tt>: signed less than</li>
5916 <li><tt>sle</tt>: signed less or equal</li>
5917</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5918
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]5919<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5920 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5921 typed. They must also be identical types.</p>
5922
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5923<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5924<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5925 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5926 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]5927 result, as follows:</p>
5928
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5929<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5930 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5931 <tt>false</tt> otherwise. No sign interpretation is necessary or
5932 performed.</li>
5933
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5934 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5935 <tt>false</tt> otherwise. No sign interpretation is necessary or
5936 performed.</li>
5937
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5938 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5939 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5940
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5941 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5942 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5943 to <tt>op2</tt>.</li>
5944
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5945 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5946 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5947
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5948 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5949 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5950
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5951 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5952 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5953
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5954 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5955 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5956 to <tt>op2</tt>.</li>
5957
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5958 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5959 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5960
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5961 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5962 <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]5963</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5964
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5965<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5966 values are compared as if they were integers.</p>
5967
5968<p>If the operands are integer vectors, then they are compared element by
5969 element. The result is an <tt>i1</tt> vector with the same number of elements
5970 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5971
5972<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5973<pre>
5974 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5975 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5976 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5977 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5978 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5979 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5980</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5981
5982<p>Note that the code generator does not yet support vector types with
5983 the <tt>icmp</tt> instruction.</p>
5984
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5985</div>
5986
5987<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5988<h4>
5989 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5990</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5991
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5992<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5993
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5994<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5995<pre>
5996 &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]5997</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5998
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5999<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6000<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
6001 values based on comparison of its operands.</p>
6002
6003<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]6004(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6005
6006<p>If the operands are floating point vectors, then the result type is a vector
6007 of boolean with the same number of elements as the operands being
6008 compared.</p>
6009
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6010<h5>Arguments:</h5>
6011<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6012 the condition code indicating the kind of comparison to perform. It is not a
6013 value, just a keyword. The possible condition code are:</p>
6014
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6015<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6016 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6017 <li><tt>oeq</tt>: ordered and equal</li>
6018 <li><tt>ogt</tt>: ordered and greater than </li>
6019 <li><tt>oge</tt>: ordered and greater than or equal</li>
6020 <li><tt>olt</tt>: ordered and less than </li>
6021 <li><tt>ole</tt>: ordered and less than or equal</li>
6022 <li><tt>one</tt>: ordered and not equal</li>
6023 <li><tt>ord</tt>: ordered (no nans)</li>
6024 <li><tt>ueq</tt>: unordered or equal</li>
6025 <li><tt>ugt</tt>: unordered or greater than </li>
6026 <li><tt>uge</tt>: unordered or greater than or equal</li>
6027 <li><tt>ult</tt>: unordered or less than </li>
6028 <li><tt>ule</tt>: unordered or less than or equal</li>
6029 <li><tt>une</tt>: unordered or not equal</li>
6030 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6031 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6032</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6033
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6034<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6035 <i>unordered</i> means that either operand may be a QNAN.</p>
6036
6037<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
6038 a <a href="#t_floating">floating point</a> type or
6039 a <a href="#t_vector">vector</a> of floating point type. They must have
6040 identical types.</p>
6041
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6042<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]6043<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6044 according to the condition code given as <tt>cond</tt>. If the operands are
6045 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]6046 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6047 follows:</p>
6048
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6049<ol>
6050 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6051
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6052 <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]6053 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
6054
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6055 <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]6056 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6057
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6058 <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]6059 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
6060
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6061 <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]6062 <tt>op1</tt> is less than <tt>op2</tt>.</li>
6063
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6064 <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]6065 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
6066
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6067 <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]6068 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
6069
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6070 <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]6071
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6072 <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]6073 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
6074
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6075 <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]6076 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
6077
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6078 <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]6079 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
6080
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6081 <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]6082 <tt>op1</tt> is less than <tt>op2</tt>.</li>
6083
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6084 <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]6085 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
6086
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6087 <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]6088 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
6089
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6090 <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]6091
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6092 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
6093</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6094
6095<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6096<pre>
6097 &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]6098 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
6099 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
6100 &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]6101</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6102
6103<p>Note that the code generator does not yet support vector types with
6104 the <tt>fcmp</tt> instruction.</p>
6105
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6106</div>
6107
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6108<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6109<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6110 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6111</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6112
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6113<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6114
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6115<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6116<pre>
6117 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
6118</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6119
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6120<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6121<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
6122 SSA graph representing the function.</p>
6123
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6124<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6125<p>The type of the incoming values is specified with the first type field. After
6126 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
6127 one pair for each predecessor basic block of the current block. Only values
6128 of <a href="#t_firstclass">first class</a> type may be used as the value
6129 arguments to the PHI node. Only labels may be used as the label
6130 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6131
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6132<p>There must be no non-phi instructions between the start of a basic block and
6133 the PHI instructions: i.e. PHI instructions must be first in a basic
6134 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6135
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6136<p>For the purposes of the SSA form, the use of each incoming value is deemed to
6137 occur on the edge from the corresponding predecessor block to the current
6138 block (but after any definition of an '<tt>invoke</tt>' instruction's return
6139 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +0000[diff] [blame]6140
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6141<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6142<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6143 specified by the pair corresponding to the predecessor basic block that
6144 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6145
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6146<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6147<pre>
6148Loop: ; Infinite loop that counts from 0 on up...
6149 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6150 %nextindvar = add i32 %indvar, 1
6151 br label %Loop
6152</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6153
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6154</div>
6155
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6156<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6157<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6158 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6159</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6160
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6161<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6162
6163<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6164<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]6165 &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>
6166
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6167 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6168</pre>
6169
6170<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6171<p>The '<tt>select</tt>' instruction is used to choose one value based on a
6172 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6173
6174
6175<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6176<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
6177 values indicating the condition, and two values of the
6178 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
6179 vectors and the condition is a scalar, then entire vectors are selected, not
6180 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6181
6182<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6183<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
6184 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6185
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6186<p>If the condition is a vector of i1, then the value arguments must be vectors
6187 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6188
6189<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6190<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]6191 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6192</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6193
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6194</div>
6195
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]6196<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6197<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6198 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6199</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6200
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6201<div>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6202
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6203<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6204<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]6205 &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]6206</pre>
6207
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6208<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6209<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6210
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6211<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6212<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6213
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6214<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6215 <li>The optional "tail" marker indicates that the callee function does not
6216 access any allocas or varargs in the caller. Note that calls may be
6217 marked "tail" even if they do not occur before
6218 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
6219 present, the function call is eligible for tail call optimization,
6220 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +0000[diff] [blame]6221 optimized into a jump</a>. The code generator may optimize calls marked
6222 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
6223 sibling call optimization</a> when the caller and callee have
6224 matching signatures, or 2) forced tail call optimization when the
6225 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6226 <ul>
6227 <li>Caller and callee both have the calling
6228 convention <tt>fastcc</tt>.</li>
6229 <li>The call is in tail position (ret immediately follows call and ret
6230 uses value of call or is void).</li>
6231 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +0000[diff] [blame]6232 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6233 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
6234 constraints are met.</a></li>
6235 </ul>
6236 </li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6237
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6238 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
6239 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6240 defaults to using C calling conventions. The calling convention of the
6241 call must match the calling convention of the target function, or else the
6242 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6243
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6244 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
6245 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
6246 '<tt>inreg</tt>' attributes are valid here.</li>
6247
6248 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
6249 type of the return value. Functions that return no value are marked
6250 <tt><a href="#t_void">void</a></tt>.</li>
6251
6252 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
6253 being invoked. The argument types must match the types implied by this
6254 signature. This type can be omitted if the function is not varargs and if
6255 the function type does not return a pointer to a function.</li>
6256
6257 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
6258 be invoked. In most cases, this is a direct function invocation, but
6259 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
6260 to function value.</li>
6261
6262 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]6263 signature argument types and parameter attributes. All arguments must be
6264 of <a href="#t_firstclass">first class</a> type. If the function
6265 signature indicates the function accepts a variable number of arguments,
6266 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6267
6268 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
6269 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
6270 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6271</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6272
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6273<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6274<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
6275 a specified function, with its incoming arguments bound to the specified
6276 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
6277 function, control flow continues with the instruction after the function
6278 call, and the return value of the function is bound to the result
6279 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6280
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6281<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6282<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]6283 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6284 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +0000[diff] [blame]6285 %X = tail call i32 @foo() <i>; yields i32</i>
6286 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
6287 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]6288
6289 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6290 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]6291 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
6292 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +0000[diff] [blame]6293 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +0000[diff] [blame]6294 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6295</pre>
6296
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6297<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +0000[diff] [blame]6298standard C99 library as being the C99 library functions, and may perform
6299optimizations or generate code for them under that assumption. This is
6300something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6301freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6302
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6303</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6304
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6305<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6306<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]6307 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6308</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6309
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6310<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6311
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6312<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6313<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6314 &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]6315</pre>
6316
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6317<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6318<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6319 the "variable argument" area of a function call. It is used to implement the
6320 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6321
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6322<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6323<p>This instruction takes a <tt>va_list*</tt> value and the type of the
6324 argument. It returns a value of the specified argument type and increments
6325 the <tt>va_list</tt> to point to the next argument. The actual type
6326 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6327
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6328<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6329<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
6330 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
6331 to the next argument. For more information, see the variable argument
6332 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6333
6334<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]6335 take a variable number of arguments, for example, the <tt>vfprintf</tt>
6336 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6337
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6338<p><tt>va_arg</tt> is an LLVM instruction instead of
6339 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
6340 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6341
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6342<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6343<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
6344
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6345<p>Note that the code generator does not yet fully support va_arg on many
6346 targets. Also, it does not currently support va_arg with aggregate types on
6347 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]6348
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6349</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6350
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6351<!-- _______________________________________________________________________ -->
6352<h4>
6353 <a name="i_landingpad">'<tt>landingpad</tt>' Instruction</a>
6354</h4>
6355
6356<div>
6357
6358<h5>Syntax:</h5>
6359<pre>
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6360 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; &lt;clause&gt;+
6361 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; cleanup &lt;clause&gt;*
Bill Wendlingbf13ee12011-08-08 08:06:05 +0000[diff] [blame]6362
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6363 &lt;clause&gt; := catch &lt;type&gt; &lt;value&gt;
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6364 &lt;clause&gt; := filter &lt;array constant type&gt; &lt;array constant&gt;
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6365</pre>
6366
6367<h5>Overview:</h5>
6368<p>The '<tt>landingpad</tt>' instruction is used by
6369 <a href="ExceptionHandling.html#overview">LLVM's exception handling
6370 system</a> to specify that a basic block is a landing pad &mdash; one where
6371 the exception lands, and corresponds to the code found in the
6372 <i><tt>catch</tt></i> portion of a <i><tt>try/catch</tt></i> sequence. It
6373 defines values supplied by the personality function (<tt>pers_fn</tt>) upon
6374 re-entry to the function. The <tt>resultval</tt> has the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6375 type <tt>resultty</tt>.</p>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6376
6377<h5>Arguments:</h5>
6378<p>This instruction takes a <tt>pers_fn</tt> value. This is the personality
6379 function associated with the unwinding mechanism. The optional
6380 <tt>cleanup</tt> flag indicates that the landing pad block is a cleanup.</p>
6381
6382<p>A <tt>clause</tt> begins with the clause type &mdash; <tt>catch</tt>
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6383 or <tt>filter</tt> &mdash; and contains the global variable representing the
6384 "type" that may be caught or filtered respectively. Unlike the
6385 <tt>catch</tt> clause, the <tt>filter</tt> clause takes an array constant as
6386 its argument. Use "<tt>[0 x i8**] undef</tt>" for a filter which cannot
6387 throw. The '<tt>landingpad</tt>' instruction must contain <em>at least</em>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6388 one <tt>clause</tt> or the <tt>cleanup</tt> flag.</p>
6389
6390<h5>Semantics:</h5>
6391<p>The '<tt>landingpad</tt>' instruction defines the values which are set by the
6392 personality function (<tt>pers_fn</tt>) upon re-entry to the function, and
6393 therefore the "result type" of the <tt>landingpad</tt> instruction. As with
6394 calling conventions, how the personality function results are represented in
6395 LLVM IR is target specific.</p>
6396
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6397<p>The clauses are applied in order from top to bottom. If two
6398 <tt>landingpad</tt> instructions are merged together through inlining, the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6399 clauses from the calling function are appended to the list of clauses.
6400 When the call stack is being unwound due to an exception being thrown, the
6401 exception is compared against each <tt>clause</tt> in turn. If it doesn't
6402 match any of the clauses, and the <tt>cleanup</tt> flag is not set, then
6403 unwinding continues further up the call stack.</p>
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6404
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6405<p>The <tt>landingpad</tt> instruction has several restrictions:</p>
6406
6407<ul>
6408 <li>A landing pad block is a basic block which is the unwind destination of an
6409 '<tt>invoke</tt>' instruction.</li>
6410 <li>A landing pad block must have a '<tt>landingpad</tt>' instruction as its
6411 first non-PHI instruction.</li>
6412 <li>There can be only one '<tt>landingpad</tt>' instruction within the landing
6413 pad block.</li>
6414 <li>A basic block that is not a landing pad block may not include a
6415 '<tt>landingpad</tt>' instruction.</li>
6416 <li>All '<tt>landingpad</tt>' instructions in a function must have the same
6417 personality function.</li>
6418</ul>
6419
6420<h5>Example:</h5>
6421<pre>
6422 ;; A landing pad which can catch an integer.
6423 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6424 catch i8** @_ZTIi
6425 ;; A landing pad that is a cleanup.
6426 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6427 cleanup
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6428 ;; A landing pad which can catch an integer and can only throw a double.
6429 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6430 catch i8** @_ZTIi
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6431 filter [1 x i8**] [@_ZTId]
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6432</pre>
6433
6434</div>
6435
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6436</div>
6437
6438</div>
6439
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6440<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6441<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6442<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6443
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6444<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6445
6446<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6447 well known names and semantics and are required to follow certain
6448 restrictions. Overall, these intrinsics represent an extension mechanism for
6449 the LLVM language that does not require changing all of the transformations
6450 in LLVM when adding to the language (or the bitcode reader/writer, the
6451 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6452
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]6453<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6454 prefix is reserved in LLVM for intrinsic names; thus, function names may not
6455 begin with this prefix. Intrinsic functions must always be external
6456 functions: you cannot define the body of intrinsic functions. Intrinsic
6457 functions may only be used in call or invoke instructions: it is illegal to
6458 take the address of an intrinsic function. Additionally, because intrinsic
6459 functions are part of the LLVM language, it is required if any are added that
6460 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6461
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6462<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6463 family of functions that perform the same operation but on different data
6464 types. Because LLVM can represent over 8 million different integer types,
6465 overloading is used commonly to allow an intrinsic function to operate on any
6466 integer type. One or more of the argument types or the result type can be
6467 overloaded to accept any integer type. Argument types may also be defined as
6468 exactly matching a previous argument's type or the result type. This allows
6469 an intrinsic function which accepts multiple arguments, but needs all of them
6470 to be of the same type, to only be overloaded with respect to a single
6471 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6472
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6473<p>Overloaded intrinsics will have the names of its overloaded argument types
6474 encoded into its function name, each preceded by a period. Only those types
6475 which are overloaded result in a name suffix. Arguments whose type is matched
6476 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6477 can take an integer of any width and returns an integer of exactly the same
6478 integer width. This leads to a family of functions such as
6479 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6480 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6481 suffix is required. Because the argument's type is matched against the return
6482 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6483
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6484<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6485 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6486
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6487<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6488<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6489 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6490</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6491
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6492<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6493
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6494<p>Variable argument support is defined in LLVM with
6495 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6496 intrinsic functions. These functions are related to the similarly named
6497 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6498
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6499<p>All of these functions operate on arguments that use a target-specific value
6500 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6501 not define what this type is, so all transformations should be prepared to
6502 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6503
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]6504<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]6505 instruction and the variable argument handling intrinsic functions are
6506 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6507
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]6508<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6509define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6510 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6511 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6512 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6513 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6514
6515 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6516 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6517
6518 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6519 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6520 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6521 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6522 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6523
6524 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6525 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]6526 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6527}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6528
6529declare void @llvm.va_start(i8*)
6530declare void @llvm.va_copy(i8*, i8*)
6531declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6532</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6533
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6534<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6535<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6536 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6537</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6538
6539
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6540<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6541
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6542<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6543<pre>
6544 declare void %llvm.va_start(i8* &lt;arglist&gt;)
6545</pre>
6546
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6547<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6548<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
6549 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]6550
6551<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6552<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]6553
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6554<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6555<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]6556 macro available in C. In a target-dependent way, it initializes
6557 the <tt>va_list</tt> element to which the argument points, so that the next
6558 call to <tt>va_arg</tt> will produce the first variable argument passed to
6559 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6560 need to know the last argument of the function as the compiler can figure
6561 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6562
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6563</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6564
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6565<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6566<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6567 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6568</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6569
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6570<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6571
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6572<h5>Syntax:</h5>
6573<pre>
6574 declare void @llvm.va_end(i8* &lt;arglist&gt;)
6575</pre>
6576
6577<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6578<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6579 which has been initialized previously
6580 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6581 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6582
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6583<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6584<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]6585
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6586<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6587<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]6588 macro available in C. In a target-dependent way, it destroys
6589 the <tt>va_list</tt> element to which the argument points. Calls
6590 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6591 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6592 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6593
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6594</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6595
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6596<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6597<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6598 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6599</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6600
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6601<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6602
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6603<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6604<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6605 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6606</pre>
6607
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6608<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6609<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6610 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6611
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6612<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6613<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]6614 The second argument is a pointer to a <tt>va_list</tt> element to copy
6615 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6616
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6617<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6618<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]6619 macro available in C. In a target-dependent way, it copies the
6620 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6621 element. This intrinsic is necessary because
6622 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6623 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6624
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6625</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6626
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6627</div>
6628
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6629<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6630<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6631 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6632</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6633
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6634<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6635
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6636<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]6637Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6638intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6639roots on the stack</a>, as well as garbage collector implementations that
6640require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6641barriers. Front-ends for type-safe garbage collected languages should generate
6642these intrinsics to make use of the LLVM garbage collectors. For more details,
6643see <a href="GarbageCollection.html">Accurate Garbage Collection with
6644LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6645
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6646<p>The garbage collection intrinsics only operate on objects in the generic
6647 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6648
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6649<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6650<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6651 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6652</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6653
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6654<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6655
6656<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6657<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6658 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6659</pre>
6660
6661<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]6662<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]6663 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6664
6665<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6666<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6667 root pointer. The second pointer (which must be either a constant or a
6668 global value address) contains the meta-data to be associated with the
6669 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6670
6671<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]6672<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]6673 location. At compile-time, the code generator generates information to allow
6674 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6675 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6676 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6677
6678</div>
6679
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6680<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6681<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6682 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6683</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6684
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6685<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6686
6687<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6688<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6689 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6690</pre>
6691
6692<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6693<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6694 locations, allowing garbage collector implementations that require read
6695 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6696
6697<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6698<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]6699 allocated from the garbage collector. The first object is a pointer to the
6700 start of the referenced object, if needed by the language runtime (otherwise
6701 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6702
6703<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6704<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6705 instruction, but may be replaced with substantially more complex code by the
6706 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6707 may only be used in a function which <a href="#gc">specifies a GC
6708 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6709
6710</div>
6711
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6712<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6713<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6714 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6715</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6716
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6717<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6718
6719<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6720<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6721 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6722</pre>
6723
6724<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6725<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6726 locations, allowing garbage collector implementations that require write
6727 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6728
6729<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6730<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]6731 object to store it to, and the third is the address of the field of Obj to
6732 store to. If the runtime does not require a pointer to the object, Obj may
6733 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6734
6735<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6736<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6737 instruction, but may be replaced with substantially more complex code by the
6738 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6739 may only be used in a function which <a href="#gc">specifies a GC
6740 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6741
6742</div>
6743
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6744</div>
6745
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6746<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6747<h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6748 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6749</h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6750
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6751<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6752
6753<p>These intrinsics are provided by LLVM to expose special features that may
6754 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6755
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6756<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6757<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6758 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6759</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6760
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6761<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6762
6763<h5>Syntax:</h5>
6764<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6765 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6766</pre>
6767
6768<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6769<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6770 target-specific value indicating the return address of the current function
6771 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6772
6773<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6774<p>The argument to this intrinsic indicates which function to return the address
6775 for. Zero indicates the calling function, one indicates its caller, etc.
6776 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6777
6778<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6779<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6780 indicating the return address of the specified call frame, or zero if it
6781 cannot be identified. The value returned by this intrinsic is likely to be
6782 incorrect or 0 for arguments other than zero, so it should only be used for
6783 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6784
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6785<p>Note that calling this intrinsic does not prevent function inlining or other
6786 aggressive transformations, so the value returned may not be that of the
6787 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6788
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6789</div>
6790
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6791<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6792<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6793 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6794</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6795
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6796<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6797
6798<h5>Syntax:</h5>
6799<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6800 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6801</pre>
6802
6803<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6804<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6805 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6806
6807<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6808<p>The argument to this intrinsic indicates which function to return the frame
6809 pointer for. Zero indicates the calling function, one indicates its caller,
6810 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6811
6812<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6813<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6814 indicating the frame address of the specified call frame, or zero if it
6815 cannot be identified. The value returned by this intrinsic is likely to be
6816 incorrect or 0 for arguments other than zero, so it should only be used for
6817 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6818
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6819<p>Note that calling this intrinsic does not prevent function inlining or other
6820 aggressive transformations, so the value returned may not be that of the
6821 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6822
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6823</div>
6824
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6825<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6826<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6827 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6828</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6829
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6830<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6831
6832<h5>Syntax:</h5>
6833<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6834 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6835</pre>
6836
6837<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6838<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6839 of the function stack, for use
6840 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6841 useful for implementing language features like scoped automatic variable
6842 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6843
6844<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6845<p>This intrinsic returns a opaque pointer value that can be passed
6846 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6847 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6848 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6849 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6850 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6851 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6852
6853</div>
6854
6855<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6856<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6857 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6858</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6859
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6860<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6861
6862<h5>Syntax:</h5>
6863<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6864 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6865</pre>
6866
6867<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6868<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6869 the function stack to the state it was in when the
6870 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6871 executed. This is useful for implementing language features like scoped
6872 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6873
6874<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6875<p>See the description
6876 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6877
6878</div>
6879
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6880<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6881<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6882 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6883</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6884
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6885<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6886
6887<h5>Syntax:</h5>
6888<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6889 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]6890</pre>
6891
6892<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6893<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6894 insert a prefetch instruction if supported; otherwise, it is a noop.
6895 Prefetches have no effect on the behavior of the program but can change its
6896 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6897
6898<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6899<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6900 specifier determining if the fetch should be for a read (0) or write (1),
6901 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6902 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6903 specifies whether the prefetch is performed on the data (1) or instruction (0)
6904 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6905 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6906
6907<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6908<p>This intrinsic does not modify the behavior of the program. In particular,
6909 prefetches cannot trap and do not produce a value. On targets that support
6910 this intrinsic, the prefetch can provide hints to the processor cache for
6911 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6912
6913</div>
6914
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6915<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6916<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6917 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6918</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6919
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6920<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6921
6922<h5>Syntax:</h5>
6923<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6924 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6925</pre>
6926
6927<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6928<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6929 Counter (PC) in a region of code to simulators and other tools. The method
6930 is target specific, but it is expected that the marker will use exported
6931 symbols to transmit the PC of the marker. The marker makes no guarantees
6932 that it will remain with any specific instruction after optimizations. It is
6933 possible that the presence of a marker will inhibit optimizations. The
6934 intended use is to be inserted after optimizations to allow correlations of
6935 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6936
6937<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6938<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6939
6940<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6941<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6942 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6943
6944</div>
6945
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6946<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6947<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6948 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6949</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6950
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6951<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6952
6953<h5>Syntax:</h5>
6954<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]6955 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6956</pre>
6957
6958<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6959<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6960 counter register (or similar low latency, high accuracy clocks) on those
6961 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6962 should map to RPCC. As the backing counters overflow quickly (on the order
6963 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6964
6965<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6966<p>When directly supported, reading the cycle counter should not modify any
6967 memory. Implementations are allowed to either return a application specific
6968 value or a system wide value. On backends without support, this is lowered
6969 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6970
6971</div>
6972
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6973</div>
6974
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6975<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6976<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6977 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6978</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6979
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6980<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6981
6982<p>LLVM provides intrinsics for a few important standard C library functions.
6983 These intrinsics allow source-language front-ends to pass information about
6984 the alignment of the pointer arguments to the code generator, providing
6985 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6986
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6987<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6988<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6989 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6990</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6991
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6992<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6993
6994<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6995<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]6996 integer bit width and for different address spaces. Not all targets support
6997 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6998
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6999<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7000 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]7001 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7002 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]7003 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7004</pre>
7005
7006<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7007<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
7008 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7009
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7010<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7011 intrinsics do not return a value, takes extra alignment/isvolatile arguments
7012 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7013
7014<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7015
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7016<p>The first argument is a pointer to the destination, the second is a pointer
7017 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7018 number of bytes to copy, the fourth argument is the alignment of the
7019 source and destination locations, and the fifth is a boolean indicating a
7020 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7021
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7022<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]7023 then the caller guarantees that both the source and destination pointers are
7024 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]7025
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7026<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7027 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
7028 The detailed access behavior is not very cleanly specified and it is unwise
7029 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7030
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7031<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7032
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7033<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
7034 source location to the destination location, which are not allowed to
7035 overlap. It copies "len" bytes of memory over. If the argument is known to
7036 be aligned to some boundary, this can be specified as the fourth argument,
7037 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7038
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7039</div>
7040
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7041<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7042<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7043 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7044</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7045
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7046<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7047
7048<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]7049<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]7050 width and for different address space. Not all targets support all bit
7051 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7052
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7053<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7054 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]7055 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7056 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]7057 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7058</pre>
7059
7060<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7061<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
7062 source location to the destination location. It is similar to the
7063 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
7064 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7065
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7066<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7067 intrinsics do not return a value, takes extra alignment/isvolatile arguments
7068 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7069
7070<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7071
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7072<p>The first argument is a pointer to the destination, the second is a pointer
7073 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7074 number of bytes to copy, the fourth argument is the alignment of the
7075 source and destination locations, and the fifth is a boolean indicating a
7076 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7077
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7078<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]7079 then the caller guarantees that the source and destination pointers are
7080 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]7081
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7082<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7083 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
7084 The detailed access behavior is not very cleanly specified and it is unwise
7085 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7086
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7087<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7088
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7089<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
7090 source location to the destination location, which may overlap. It copies
7091 "len" bytes of memory over. If the argument is known to be aligned to some
7092 boundary, this can be specified as the fourth argument, otherwise it should
7093 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7094
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7095</div>
7096
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7097<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7098<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7099 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7100</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7101
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7102<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7103
7104<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]7105<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]7106 width and for different address spaces. However, not all targets support all
7107 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7108
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7109<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7110 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]7111 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7112 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]7113 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7114</pre>
7115
7116<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7117<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
7118 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7119
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7120<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]7121 intrinsic does not return a value and takes extra alignment/volatile
7122 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7123
7124<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7125<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]7126 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7127 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]7128 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7129
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7130<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]7131 then the caller guarantees that the destination pointer is aligned to that
7132 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7133
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7134<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7135 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
7136 The detailed access behavior is not very cleanly specified and it is unwise
7137 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7138
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7139<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7140<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
7141 at the destination location. If the argument is known to be aligned to some
7142 boundary, this can be specified as the fourth argument, otherwise it should
7143 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7144
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7145</div>
7146
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7147<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7148<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7149 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7150</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7151
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7152<div>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7153
7154<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7155<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
7156 floating point or vector of floating point type. Not all targets support all
7157 types however.</p>
7158
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7159<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]7160 declare float @llvm.sqrt.f32(float %Val)
7161 declare double @llvm.sqrt.f64(double %Val)
7162 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7163 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7164 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7165</pre>
7166
7167<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7168<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
7169 returning the same value as the libm '<tt>sqrt</tt>' functions would.
7170 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
7171 behavior for negative numbers other than -0.0 (which allows for better
7172 optimization, because there is no need to worry about errno being
7173 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]7174
7175<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7176<p>The argument and return value are floating point numbers of the same
7177 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7178
7179<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7180<p>This function returns the sqrt of the specified operand if it is a
7181 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7182
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7183</div>
7184
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7185<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7186<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7187 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7188</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7189
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7190<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7191
7192<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7193<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
7194 floating point or vector of floating point type. Not all targets support all
7195 types however.</p>
7196
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7197<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]7198 declare float @llvm.powi.f32(float %Val, i32 %power)
7199 declare double @llvm.powi.f64(double %Val, i32 %power)
7200 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7201 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7202 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7203</pre>
7204
7205<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7206<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
7207 specified (positive or negative) power. The order of evaluation of
7208 multiplications is not defined. When a vector of floating point type is
7209 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7210
7211<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7212<p>The second argument is an integer power, and the first is a value to raise to
7213 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7214
7215<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7216<p>This function returns the first value raised to the second power with an
7217 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7218
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7219</div>
7220
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7221<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7222<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7223 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7224</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7225
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7226<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7227
7228<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7229<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
7230 floating point or vector of floating point type. Not all targets support all
7231 types however.</p>
7232
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7233<pre>
7234 declare float @llvm.sin.f32(float %Val)
7235 declare double @llvm.sin.f64(double %Val)
7236 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7237 declare fp128 @llvm.sin.f128(fp128 %Val)
7238 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7239</pre>
7240
7241<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7242<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7243
7244<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7245<p>The argument and return value are floating point numbers of the same
7246 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7247
7248<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7249<p>This function returns the sine of the specified operand, returning the same
7250 values as the libm <tt>sin</tt> functions would, and handles error conditions
7251 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7252
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7253</div>
7254
7255<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7256<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7257 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7258</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7259
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7260<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7261
7262<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7263<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
7264 floating point or vector of floating point type. Not all targets support all
7265 types however.</p>
7266
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7267<pre>
7268 declare float @llvm.cos.f32(float %Val)
7269 declare double @llvm.cos.f64(double %Val)
7270 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7271 declare fp128 @llvm.cos.f128(fp128 %Val)
7272 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7273</pre>
7274
7275<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7276<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7277
7278<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7279<p>The argument and return value are floating point numbers of the same
7280 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7281
7282<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7283<p>This function returns the cosine of the specified operand, returning the same
7284 values as the libm <tt>cos</tt> functions would, and handles error conditions
7285 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7286
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7287</div>
7288
7289<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7290<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7291 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7292</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7293
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7294<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7295
7296<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7297<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
7298 floating point or vector of floating point type. Not all targets support all
7299 types however.</p>
7300
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7301<pre>
7302 declare float @llvm.pow.f32(float %Val, float %Power)
7303 declare double @llvm.pow.f64(double %Val, double %Power)
7304 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7305 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7306 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7307</pre>
7308
7309<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7310<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
7311 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7312
7313<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7314<p>The second argument is a floating point power, and the first is a value to
7315 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7316
7317<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7318<p>This function returns the first value raised to the second power, returning
7319 the same values as the libm <tt>pow</tt> functions would, and handles error
7320 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7321
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7322</div>
7323
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7324<!-- _______________________________________________________________________ -->
7325<h4>
7326 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
7327</h4>
7328
7329<div>
7330
7331<h5>Syntax:</h5>
7332<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
7333 floating point or vector of floating point type. Not all targets support all
7334 types however.</p>
7335
7336<pre>
7337 declare float @llvm.exp.f32(float %Val)
7338 declare double @llvm.exp.f64(double %Val)
7339 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7340 declare fp128 @llvm.exp.f128(fp128 %Val)
7341 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7342</pre>
7343
7344<h5>Overview:</h5>
7345<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
7346
7347<h5>Arguments:</h5>
7348<p>The argument and return value are floating point numbers of the same
7349 type.</p>
7350
7351<h5>Semantics:</h5>
7352<p>This function returns the same values as the libm <tt>exp</tt> functions
7353 would, and handles error conditions in the same way.</p>
7354
7355</div>
7356
7357<!-- _______________________________________________________________________ -->
7358<h4>
7359 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
7360</h4>
7361
7362<div>
7363
7364<h5>Syntax:</h5>
7365<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
7366 floating point or vector of floating point type. Not all targets support all
7367 types however.</p>
7368
7369<pre>
7370 declare float @llvm.log.f32(float %Val)
7371 declare double @llvm.log.f64(double %Val)
7372 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7373 declare fp128 @llvm.log.f128(fp128 %Val)
7374 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7375</pre>
7376
7377<h5>Overview:</h5>
7378<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
7379
7380<h5>Arguments:</h5>
7381<p>The argument and return value are floating point numbers of the same
7382 type.</p>
7383
7384<h5>Semantics:</h5>
7385<p>This function returns the same values as the libm <tt>log</tt> functions
7386 would, and handles error conditions in the same way.</p>
7387
Nick Lewycky1c929be2011-10-31 01:32:21 +0000[diff] [blame]7388</div>
7389
7390<!-- _______________________________________________________________________ -->
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7391<h4>
7392 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
7393</h4>
7394
7395<div>
7396
7397<h5>Syntax:</h5>
7398<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
7399 floating point or vector of floating point type. Not all targets support all
7400 types however.</p>
7401
7402<pre>
7403 declare float @llvm.fma.f32(float %a, float %b, float %c)
7404 declare double @llvm.fma.f64(double %a, double %b, double %c)
7405 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7406 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7407 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7408</pre>
7409
7410<h5>Overview:</h5>
Cameron Zwarichabc43e62011-07-08 22:13:55 +0000[diff] [blame]7411<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7412 operation.</p>
7413
7414<h5>Arguments:</h5>
7415<p>The argument and return value are floating point numbers of the same
7416 type.</p>
7417
7418<h5>Semantics:</h5>
7419<p>This function returns the same values as the libm <tt>fma</tt> functions
7420 would.</p>
7421
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7422</div>
7423
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]7424</div>
7425
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7426<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7427<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7428 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7429</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7430
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7431<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7432
7433<p>LLVM provides intrinsics for a few important bit manipulation operations.
7434 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7435
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7436<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7437<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7438 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7439</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7440
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7441<div>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7442
7443<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7444<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7445 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
7446
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7447<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7448 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
7449 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
7450 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7451</pre>
7452
7453<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7454<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
7455 values with an even number of bytes (positive multiple of 16 bits). These
7456 are useful for performing operations on data that is not in the target's
7457 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7458
7459<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7460<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
7461 and low byte of the input i16 swapped. Similarly,
7462 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
7463 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7464 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7465 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7466 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7467 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7468
7469</div>
7470
7471<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7472<h4>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]7473 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7474</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7475
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7476<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7477
7478<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7479<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]7480 width, or on any vector with integer elements. Not all targets support all
7481 bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7482
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7483<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7484 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7485 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]7486 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7487 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
7488 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7489 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]7490</pre>
7491
7492<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7493<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7494 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7495
7496<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7497<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]7498 integer type, or a vector with integer elements.
7499 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7500
7501<h5>Semantics:</h5>
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7502<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7503 element of a vector.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7504
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7505</div>
7506
7507<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7508<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7509 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7510</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7511
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7512<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7513
7514<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7515<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]7516 integer bit width, or any vector whose elements are integers. Not all
7517 targets support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7518
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7519<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7520 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7521 declare i16 @llvm.ctlz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7522 declare i32 @llvm.ctlz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7523 declare i64 @llvm.ctlz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7524 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7525 declase &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7526</pre>
7527
7528<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7529<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7530 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7531
7532<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7533<p>The first argument is the value to be counted. This argument may be of any
7534 integer type, or a vectory with integer element type. The return type
7535 must match the first argument type.</p>
7536
7537<p>The second argument must be a constant and is a flag to indicate whether the
7538 intrinsic should ensure that a zero as the first argument produces a defined
7539 result. Historically some architectures did not provide a defined result for
7540 zero values as efficiently, and many algorithms are now predicated on
7541 avoiding zero-value inputs.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7542
7543<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7544<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7545 zeros in a variable, or within each element of the vector.
7546 If <tt>src == 0</tt> then the result is the size in bits of the type of
7547 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7548 For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7549
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7550</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7551
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7552<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7553<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7554 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7555</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7556
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7557<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7558
7559<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7560<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]7561 integer bit width, or any vector of integer elements. Not all targets
7562 support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7563
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7564<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7565 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7566 declare i16 @llvm.cttz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7567 declare i32 @llvm.cttz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7568 declare i64 @llvm.cttz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7569 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7570 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7571</pre>
7572
7573<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7574<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7575 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7576
7577<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7578<p>The first argument is the value to be counted. This argument may be of any
7579 integer type, or a vectory with integer element type. The return type
7580 must match the first argument type.</p>
7581
7582<p>The second argument must be a constant and is a flag to indicate whether the
7583 intrinsic should ensure that a zero as the first argument produces a defined
7584 result. Historically some architectures did not provide a defined result for
7585 zero values as efficiently, and many algorithms are now predicated on
7586 avoiding zero-value inputs.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7587
7588<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7589<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7590 zeros in a variable, or within each element of a vector.
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7591 If <tt>src == 0</tt> then the result is the size in bits of the type of
7592 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7593 For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7594
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7595</div>
7596
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7597</div>
7598
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7599<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7600<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7601 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7602</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7603
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7604<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7605
7606<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7607
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7608<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7609<h4>
7610 <a name="int_sadd_overflow">
7611 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7612 </a>
7613</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7614
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7615<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7616
7617<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7618<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]7619 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7620
7621<pre>
7622 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7623 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7624 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7625</pre>
7626
7627<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7628<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7629 a signed addition of the two arguments, and indicate whether an overflow
7630 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7631
7632<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7633<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]7634 be of integer types of any bit width, but they must have the same bit
7635 width. The second element of the result structure must be of
7636 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7637 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7638
7639<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7640<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7641 a signed addition of the two variables. They return a structure &mdash; the
7642 first element of which is the signed summation, and the second element of
7643 which is a bit specifying if the signed summation resulted in an
7644 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7645
7646<h5>Examples:</h5>
7647<pre>
7648 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7649 %sum = extractvalue {i32, i1} %res, 0
7650 %obit = extractvalue {i32, i1} %res, 1
7651 br i1 %obit, label %overflow, label %normal
7652</pre>
7653
7654</div>
7655
7656<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7657<h4>
7658 <a name="int_uadd_overflow">
7659 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7660 </a>
7661</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7662
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7663<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7664
7665<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7666<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]7667 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7668
7669<pre>
7670 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7671 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7672 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7673</pre>
7674
7675<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7676<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7677 an unsigned addition of the two arguments, and indicate whether a carry
7678 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7679
7680<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7681<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]7682 be of integer types of any bit width, but they must have the same bit
7683 width. The second element of the result structure must be of
7684 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7685 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7686
7687<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7688<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7689 an unsigned addition of the two arguments. They return a structure &mdash;
7690 the first element of which is the sum, and the second element of which is a
7691 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7692
7693<h5>Examples:</h5>
7694<pre>
7695 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7696 %sum = extractvalue {i32, i1} %res, 0
7697 %obit = extractvalue {i32, i1} %res, 1
7698 br i1 %obit, label %carry, label %normal
7699</pre>
7700
7701</div>
7702
7703<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7704<h4>
7705 <a name="int_ssub_overflow">
7706 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7707 </a>
7708</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7709
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7710<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7711
7712<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7713<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]7714 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7715
7716<pre>
7717 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7718 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7719 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7720</pre>
7721
7722<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7723<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7724 a signed subtraction of the two arguments, and indicate whether an overflow
7725 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7726
7727<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7728<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]7729 be of integer types of any bit width, but they must have the same bit
7730 width. The second element of the result structure must be of
7731 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7732 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7733
7734<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7735<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7736 a signed subtraction of the two arguments. They return a structure &mdash;
7737 the first element of which is the subtraction, and the second element of
7738 which is a bit specifying if the signed subtraction resulted in an
7739 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7740
7741<h5>Examples:</h5>
7742<pre>
7743 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7744 %sum = extractvalue {i32, i1} %res, 0
7745 %obit = extractvalue {i32, i1} %res, 1
7746 br i1 %obit, label %overflow, label %normal
7747</pre>
7748
7749</div>
7750
7751<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7752<h4>
7753 <a name="int_usub_overflow">
7754 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7755 </a>
7756</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7757
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7758<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7759
7760<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7761<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]7762 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7763
7764<pre>
7765 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7766 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7767 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7768</pre>
7769
7770<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7771<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7772 an unsigned subtraction of the two arguments, and indicate whether an
7773 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7774
7775<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7776<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]7777 be of integer types of any bit width, but they must have the same bit
7778 width. The second element of the result structure must be of
7779 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7780 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7781
7782<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7783<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7784 an unsigned subtraction of the two arguments. They return a structure &mdash;
7785 the first element of which is the subtraction, and the second element of
7786 which is a bit specifying if the unsigned subtraction resulted in an
7787 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7788
7789<h5>Examples:</h5>
7790<pre>
7791 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7792 %sum = extractvalue {i32, i1} %res, 0
7793 %obit = extractvalue {i32, i1} %res, 1
7794 br i1 %obit, label %overflow, label %normal
7795</pre>
7796
7797</div>
7798
7799<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7800<h4>
7801 <a name="int_smul_overflow">
7802 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7803 </a>
7804</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7805
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7806<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7807
7808<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7809<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]7810 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7811
7812<pre>
7813 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7814 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7815 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7816</pre>
7817
7818<h5>Overview:</h5>
7819
7820<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7821 a signed multiplication of the two arguments, and indicate whether an
7822 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7823
7824<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7825<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]7826 be of integer types of any bit width, but they must have the same bit
7827 width. The second element of the result structure must be of
7828 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7829 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7830
7831<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7832<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7833 a signed multiplication of the two arguments. They return a structure &mdash;
7834 the first element of which is the multiplication, and the second element of
7835 which is a bit specifying if the signed multiplication resulted in an
7836 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7837
7838<h5>Examples:</h5>
7839<pre>
7840 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7841 %sum = extractvalue {i32, i1} %res, 0
7842 %obit = extractvalue {i32, i1} %res, 1
7843 br i1 %obit, label %overflow, label %normal
7844</pre>
7845
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]7846</div>
7847
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7848<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7849<h4>
7850 <a name="int_umul_overflow">
7851 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7852 </a>
7853</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7854
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7855<div>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7856
7857<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7858<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]7859 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7860
7861<pre>
7862 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7863 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7864 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7865</pre>
7866
7867<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7868<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7869 a unsigned multiplication of the two arguments, and indicate whether an
7870 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7871
7872<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7873<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]7874 be of integer types of any bit width, but they must have the same bit
7875 width. The second element of the result structure must be of
7876 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7877 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7878
7879<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7880<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7881 an unsigned multiplication of the two arguments. They return a structure
7882 &mdash; the first element of which is the multiplication, and the second
7883 element of which is a bit specifying if the unsigned multiplication resulted
7884 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7885
7886<h5>Examples:</h5>
7887<pre>
7888 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7889 %sum = extractvalue {i32, i1} %res, 0
7890 %obit = extractvalue {i32, i1} %res, 1
7891 br i1 %obit, label %overflow, label %normal
7892</pre>
7893
7894</div>
7895
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7896</div>
7897
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7898<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7899<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7900 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7901</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7902
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7903<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7904
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7905<p>Half precision floating point is a storage-only format. This means that it is
7906 a dense encoding (in memory) but does not support computation in the
7907 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7908
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7909<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7910 value as an i16, then convert it to float with <a
7911 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7912 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7913 double etc). To store the value back to memory, it is first converted to
7914 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7915 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7916 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7917
7918<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7919<h4>
7920 <a name="int_convert_to_fp16">
7921 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7922 </a>
7923</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7924
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7925<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7926
7927<h5>Syntax:</h5>
7928<pre>
7929 declare i16 @llvm.convert.to.fp16(f32 %a)
7930</pre>
7931
7932<h5>Overview:</h5>
7933<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7934 a conversion from single precision floating point format to half precision
7935 floating point format.</p>
7936
7937<h5>Arguments:</h5>
7938<p>The intrinsic function contains single argument - the value to be
7939 converted.</p>
7940
7941<h5>Semantics:</h5>
7942<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7943 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7944 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7945 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7946
7947<h5>Examples:</h5>
7948<pre>
7949 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7950 store i16 %res, i16* @x, align 2
7951</pre>
7952
7953</div>
7954
7955<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7956<h4>
7957 <a name="int_convert_from_fp16">
7958 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7959 </a>
7960</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7961
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7962<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7963
7964<h5>Syntax:</h5>
7965<pre>
7966 declare f32 @llvm.convert.from.fp16(i16 %a)
7967</pre>
7968
7969<h5>Overview:</h5>
7970<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7971 a conversion from half precision floating point format to single precision
7972 floating point format.</p>
7973
7974<h5>Arguments:</h5>
7975<p>The intrinsic function contains single argument - the value to be
7976 converted.</p>
7977
7978<h5>Semantics:</h5>
7979<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7980 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7981 precision floating point format. The input half-float value is represented by
7982 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7983
7984<h5>Examples:</h5>
7985<pre>
7986 %a = load i16* @x, align 2
7987 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7988</pre>
7989
7990</div>
7991
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7992</div>
7993
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7994<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7995<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7996 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7997</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7998
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7999<div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8000
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8001<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
8002 prefix), are described in
8003 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
8004 Level Debugging</a> document.</p>
8005
8006</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8007
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8008<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8009<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8010 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8011</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8012
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8013<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8014
8015<p>The LLVM exception handling intrinsics (which all start with
8016 <tt>llvm.eh.</tt> prefix), are described in
8017 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
8018 Handling</a> document.</p>
8019
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8020</div>
8021
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8022<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8023<h3>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8024 <a name="int_trampoline">Trampoline Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8025</h3>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8026
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8027<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8028
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8029<p>These intrinsics make it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]8030 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
8031 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8032 function pointer lacking the nest parameter - the caller does not need to
8033 provide a value for it. Instead, the value to use is stored in advance in a
8034 "trampoline", a block of memory usually allocated on the stack, which also
8035 contains code to splice the nest value into the argument list. This is used
8036 to implement the GCC nested function address extension.</p>
8037
8038<p>For example, if the function is
8039 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
8040 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
8041 follows:</p>
8042
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]8043<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]8044 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8045 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8046 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8047 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]8048 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8049</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8050
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8051<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
8052 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8053
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8054<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8055<h4>
8056 <a name="int_it">
8057 '<tt>llvm.init.trampoline</tt>' Intrinsic
8058 </a>
8059</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8060
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8061<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8062
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8063<h5>Syntax:</h5>
8064<pre>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8065 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]8066</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8067
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8068<h5>Overview:</h5>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8069<p>This fills the memory pointed to by <tt>tramp</tt> with executable code,
8070 turning it into a trampoline.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8071
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8072<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8073<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
8074 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
8075 sufficiently aligned block of memory; this memory is written to by the
8076 intrinsic. Note that the size and the alignment are target-specific - LLVM
8077 currently provides no portable way of determining them, so a front-end that
8078 generates this intrinsic needs to have some target-specific knowledge.
8079 The <tt>func</tt> argument must hold a function bitcast to
8080 an <tt>i8*</tt>.</p>
8081
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8082<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8083<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]8084 dependent code, turning it into a function. Then <tt>tramp</tt> needs to be
8085 passed to <a href="#int_at">llvm.adjust.trampoline</a> to get a pointer
8086 which can be <a href="#int_trampoline">bitcast (to a new function) and
8087 called</a>. The new function's signature is the same as that of
8088 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
8089 removed. At most one such <tt>nest</tt> argument is allowed, and it must be of
8090 pointer type. Calling the new function is equivalent to calling <tt>func</tt>
8091 with the same argument list, but with <tt>nval</tt> used for the missing
8092 <tt>nest</tt> argument. If, after calling <tt>llvm.init.trampoline</tt>, the
8093 memory pointed to by <tt>tramp</tt> is modified, then the effect of any later call
8094 to the returned function pointer is undefined.</p>
8095</div>
8096
8097<!-- _______________________________________________________________________ -->
8098<h4>
8099 <a name="int_at">
8100 '<tt>llvm.adjust.trampoline</tt>' Intrinsic
8101 </a>
8102</h4>
8103
8104<div>
8105
8106<h5>Syntax:</h5>
8107<pre>
8108 declare i8* @llvm.adjust.trampoline(i8* &lt;tramp&gt;)
8109</pre>
8110
8111<h5>Overview:</h5>
8112<p>This performs any required machine-specific adjustment to the address of a
8113 trampoline (passed as <tt>tramp</tt>).</p>
8114
8115<h5>Arguments:</h5>
8116<p><tt>tramp</tt> must point to a block of memory which already has trampoline code
8117 filled in by a previous call to <a href="#int_it"><tt>llvm.init.trampoline</tt>
8118 </a>.</p>
8119
8120<h5>Semantics:</h5>
8121<p>On some architectures the address of the code to be executed needs to be
8122 different to the address where the trampoline is actually stored. This
8123 intrinsic returns the executable address corresponding to <tt>tramp</tt>
8124 after performing the required machine specific adjustments.
8125 The pointer returned can then be <a href="#int_trampoline"> bitcast and
8126 executed</a>.
8127</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8128
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8129</div>
8130
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8131</div>
8132
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8133<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8134<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8135 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8136</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8137
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8138<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8139
8140<p>This class of intrinsics exists to information about the lifetime of memory
8141 objects and ranges where variables are immutable.</p>
8142
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8143<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8144<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8145 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8146</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8147
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8148<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8149
8150<h5>Syntax:</h5>
8151<pre>
8152 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8153</pre>
8154
8155<h5>Overview:</h5>
8156<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
8157 object's lifetime.</p>
8158
8159<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8160<p>The first argument is a constant integer representing the size of the
8161 object, or -1 if it is variable sized. The second argument is a pointer to
8162 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8163
8164<h5>Semantics:</h5>
8165<p>This intrinsic indicates that before this point in the code, the value of the
8166 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]8167 never be used and has an undefined value. A load from the pointer that
8168 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8169 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
8170
8171</div>
8172
8173<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8174<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8175 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8176</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8177
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8178<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8179
8180<h5>Syntax:</h5>
8181<pre>
8182 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8183</pre>
8184
8185<h5>Overview:</h5>
8186<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
8187 object's lifetime.</p>
8188
8189<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8190<p>The first argument is a constant integer representing the size of the
8191 object, or -1 if it is variable sized. The second argument is a pointer to
8192 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8193
8194<h5>Semantics:</h5>
8195<p>This intrinsic indicates that after this point in the code, the value of the
8196 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8197 never be used and has an undefined value. Any stores into the memory object
8198 following this intrinsic may be removed as dead.
8199
8200</div>
8201
8202<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8203<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8204 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8205</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8206
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8207<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8208
8209<h5>Syntax:</h5>
8210<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +0000[diff] [blame]8211 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8212</pre>
8213
8214<h5>Overview:</h5>
8215<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8216 a memory object will not change.</p>
8217
8218<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8219<p>The first argument is a constant integer representing the size of the
8220 object, or -1 if it is variable sized. The second argument is a pointer to
8221 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8222
8223<h5>Semantics:</h5>
8224<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8225 the return value, the referenced memory location is constant and
8226 unchanging.</p>
8227
8228</div>
8229
8230<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8231<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8232 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8233</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8234
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8235<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8236
8237<h5>Syntax:</h5>
8238<pre>
8239 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8240</pre>
8241
8242<h5>Overview:</h5>
8243<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8244 a memory object are mutable.</p>
8245
8246<h5>Arguments:</h5>
8247<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8248 The second argument is a constant integer representing the size of the
8249 object, or -1 if it is variable sized and the third argument is a pointer
8250 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8251
8252<h5>Semantics:</h5>
8253<p>This intrinsic indicates that the memory is mutable again.</p>
8254
8255</div>
8256
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8257</div>
8258
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]8259<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8260<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8261 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8262</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8263
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8264<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8265
8266<p>This class of intrinsics is designed to be generic and has no specific
8267 purpose.</p>
8268
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8269<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8270<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8271 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8272</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8273
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8274<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8275
8276<h5>Syntax:</h5>
8277<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8278 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]8279</pre>
8280
8281<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8282<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8283
8284<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8285<p>The first argument is a pointer to a value, the second is a pointer to a
8286 global string, the third is a pointer to a global string which is the source
8287 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8288
8289<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8290<p>This intrinsic allows annotation of local variables with arbitrary strings.
8291 This can be useful for special purpose optimizations that want to look for
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8292 these annotations. These have no other defined use; they are ignored by code
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8293 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8294
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8295</div>
8296
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8297<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8298<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]8299 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8300</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8301
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8302<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8303
8304<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8305<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8306 any integer bit width.</p>
8307
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8308<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8309 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8310 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8311 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8312 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8313 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]8314</pre>
8315
8316<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8317<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8318
8319<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8320<p>The first argument is an integer value (result of some expression), the
8321 second is a pointer to a global string, the third is a pointer to a global
8322 string which is the source file name, and the last argument is the line
8323 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8324
8325<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8326<p>This intrinsic allows annotations to be put on arbitrary expressions with
8327 arbitrary strings. This can be useful for special purpose optimizations that
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8328 want to look for these annotations. These have no other defined use; they
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8329 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8330
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8331</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8332
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8333<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8334<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8335 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8336</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8337
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8338<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8339
8340<h5>Syntax:</h5>
8341<pre>
8342 declare void @llvm.trap()
8343</pre>
8344
8345<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8346<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8347
8348<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8349<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8350
8351<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8352<p>This intrinsics is lowered to the target dependent trap instruction. If the
8353 target does not have a trap instruction, this intrinsic will be lowered to
8354 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8355
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8356</div>
8357
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8358<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8359<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +0000[diff] [blame]8360 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8361</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8362
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8363<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8364
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8365<h5>Syntax:</h5>
8366<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8367 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8368</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8369
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8370<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8371<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8372 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8373 ensure that it is placed on the stack before local variables.</p>
8374
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8375<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8376<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8377 arguments. The first argument is the value loaded from the stack
8378 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8379 that has enough space to hold the value of the guard.</p>
8380
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8381<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8382<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8383 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8384 stack. This is to ensure that if a local variable on the stack is
8385 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8386 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8387 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8388 function.</p>
8389
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8390</div>
8391
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8392<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8393<h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8394 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8395</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8396
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8397<div>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8398
8399<h5>Syntax:</h5>
8400<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8401 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8402 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8403</pre>
8404
8405<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8406<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8407 the optimizers to determine at compile time whether a) an operation (like
8408 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8409 runtime check for overflow isn't necessary. An object in this context means
8410 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8411
8412<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8413<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8414 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8415 is a boolean 0 or 1. This argument determines whether you want the
8416 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]8417 1, variables are not allowed.</p>
8418
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8419<h5>Semantics:</h5>
8420<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8421 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8422 depending on the <tt>type</tt> argument, if the size cannot be determined at
8423 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8424
8425</div>
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8426<!-- _______________________________________________________________________ -->
8427<h4>
8428 <a name="int_expect">'<tt>llvm.expect</tt>' Intrinsic</a>
8429</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8430
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8431<div>
8432
8433<h5>Syntax:</h5>
8434<pre>
8435 declare i32 @llvm.expect.i32(i32 &lt;val&gt;, i32 &lt;expected_val&gt;)
8436 declare i64 @llvm.expect.i64(i64 &lt;val&gt;, i64 &lt;expected_val&gt;)
8437</pre>
8438
8439<h5>Overview:</h5>
8440<p>The <tt>llvm.expect</tt> intrinsic provides information about expected (the
8441 most probable) value of <tt>val</tt>, which can be used by optimizers.</p>
8442
8443<h5>Arguments:</h5>
8444<p>The <tt>llvm.expect</tt> intrinsic takes two arguments. The first
8445 argument is a value. The second argument is an expected value, this needs to
8446 be a constant value, variables are not allowed.</p>
8447
8448<h5>Semantics:</h5>
8449<p>This intrinsic is lowered to the <tt>val</tt>.</p>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8450</div>
8451
8452</div>
8453
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8454</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8455<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8456<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8457<address>
8458 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8459 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8460 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8461 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]8462
8463 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]8464 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8465 Last modified: $Date$
8466</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]8467
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]8468</body>
8469</html>