Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 0d87ec21d79c8622733b8367aa41067169602480 / . / docs / LangRef.html
blob: 395c9bbabf170cce1665aa8c926883939c432c5b [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]14
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]15<h1>LLVM Language Reference Manual</h1>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]25 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]28 <li><a href="#linkage_linker_private_weak_def_auto">'<tt>linker_private_weak_def_auto</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]29 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
30 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
31 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
32 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
33 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
34 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
35 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner5a2d8752009-10-10 18:26:06 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]37 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
38 <li><a href="#linkage_external">'<tt>externally visible</tt>' Linkage</a></li>
39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]41 </ol>
42 </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]43 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]44 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]45 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]46 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]47 <li><a href="#aliasstructure">Aliases</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]49 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]50 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]51 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]53 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]56 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]57 </ol>
58 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]59 <li><a href="#typesystem">Type System</a>
60 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]61 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]62 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]63 <ol>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]64 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]65 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]66 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]67 <li><a href="#t_void">Void Type</a></li>
68 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]69 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]70 </ol>
71 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]72 <li><a href="#t_derived">Derived Types</a>
73 <ol>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]74 <li><a href="#t_aggregate">Aggregate Types</a>
75 <ol>
76 <li><a href="#t_array">Array Type</a></li>
77 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]78 <li><a href="#t_opaque">Opaque Structure Types</a></li>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]79 <li><a href="#t_vector">Vector Type</a></li>
80 </ol>
81 </li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]82 <li><a href="#t_function">Function Type</a></li>
83 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]84 </ol>
85 </li>
86 </ol>
87 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]88 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]89 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]90 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]91 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]92 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
93 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]94 <li><a href="#trapvalues">Trap Values</a></li>
Chris Lattnerf9d078e2009-10-27 21:19:13 +0000[diff] [blame]95 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]96 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]97 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]98 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]99 <li><a href="#othervalues">Other Values</a>
100 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]101 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]102 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a></li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]103 </ol>
104 </li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]105 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
106 <ol>
107 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]108 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
109 Global Variable</a></li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]110 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
111 Global Variable</a></li>
112 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
113 Global Variable</a></li>
114 </ol>
115 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]116 <li><a href="#instref">Instruction Reference</a>
117 <ol>
118 <li><a href="#terminators">Terminator Instructions</a>
119 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]120 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
121 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]122 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]123 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]124 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]125 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]126 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]127 </ol>
128 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]129 <li><a href="#binaryops">Binary Operations</a>
130 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]131 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]132 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]133 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]134 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]135 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]136 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]137 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
138 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
139 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]140 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
141 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
142 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]143 </ol>
144 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]145 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
146 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]147 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
148 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
149 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]150 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]151 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]152 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]153 </ol>
154 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]155 <li><a href="#vectorops">Vector Operations</a>
156 <ol>
157 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
158 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
159 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]160 </ol>
161 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]162 <li><a href="#aggregateops">Aggregate Operations</a>
163 <ol>
164 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
165 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
166 </ol>
167 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]168 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]169 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]170 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]171 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
172 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]173 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]174 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]175 </ol>
176 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]177 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]178 <ol>
179 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
180 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
181 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
182 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
183 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]184 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
185 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
186 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
187 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]188 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
189 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]190 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]191 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]192 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]193 <li><a href="#otherops">Other Operations</a>
194 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]195 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
196 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]197 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]198 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]199 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]200 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]201 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]202 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]203 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]204 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]205 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]206 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]207 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
208 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]209 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
210 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
211 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]212 </ol>
213 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]214 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
215 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]216 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
217 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
218 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]219 </ol>
220 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]221 <li><a href="#int_codegen">Code Generator Intrinsics</a>
222 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]223 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
224 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
225 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
226 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
227 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
228 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000[diff] [blame]229 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]230 </ol>
231 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]232 <li><a href="#int_libc">Standard C Library Intrinsics</a>
233 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]234 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
235 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
236 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
237 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
238 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]239 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
240 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
241 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohman08b280b2011-05-27 00:36:31 +0000[diff] [blame]242 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
243 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]244 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]245 </ol>
246 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]247 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]248 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]249 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]250 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
251 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
252 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]253 </ol>
254 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]255 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
256 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]257 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
258 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
259 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
260 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
261 <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]262 <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]263 </ol>
264 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]265 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
266 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]267 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
268 <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]269 </ol>
270 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]271 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]272 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]273 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]274 <ol>
275 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]276 </ol>
277 </li>
Bill Wendling3c44f5b2008-11-18 22:10:53 +0000[diff] [blame]278 <li><a href="#int_atomics">Atomic intrinsics</a>
279 <ol>
280 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
281 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
282 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
283 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
284 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
285 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
286 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
287 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
288 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
289 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
290 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
291 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
292 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
293 </ol>
294 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]295 <li><a href="#int_memorymarkers">Memory Use Markers</a>
296 <ol>
297 <li><a href="#int_lifetime_start"><tt>llvm.lifetime.start</tt></a></li>
298 <li><a href="#int_lifetime_end"><tt>llvm.lifetime.end</tt></a></li>
299 <li><a href="#int_invariant_start"><tt>llvm.invariant.start</tt></a></li>
300 <li><a href="#int_invariant_end"><tt>llvm.invariant.end</tt></a></li>
301 </ol>
302 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]303 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]304 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]305 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]306 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]307 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]308 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]309 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]310 '<tt>llvm.trap</tt>' Intrinsic</a></li>
311 <li><a href="#int_stackprotector">
312 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]313 <li><a href="#int_objectsize">
314 '<tt>llvm.objectsize</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
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]644 <dt><tt><b><a name="linkage_external">externally visible</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
677 other than "externally visible", <tt>dllimport</tt>
678 or <tt>extern_weak</tt>.</p>
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
941 be significant and two identical functions can be merged</p>.
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>
Charles Davis1e063d12010-02-12 00:31:15 +0000[diff] [blame]1150 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1151 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1152 the backend should forcibly align the stack pointer. Specify the
1153 desired alignment, which must be a power of two, in parentheses.
1154
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1155 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1156 <dd>This attribute indicates that the inliner should attempt to inline this
1157 function into callers whenever possible, ignoring any active inlining size
1158 threshold for this caller.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1159
Charles Davis970bfcc2010-10-25 15:37:09 +0000[diff] [blame]1160 <dt><tt><b>hotpatch</b></tt></dt>
Charles Davis6f12e292010-10-25 16:29:03 +0000[diff] [blame]1161 <dd>This attribute indicates that the function should be 'hotpatchable',
Charles Davis0076d202010-10-25 19:07:39 +0000[diff] [blame]1162 meaning the function can be patched and/or hooked even while it is
1163 loaded into memory. On x86, the function prologue will be preceded
1164 by six bytes of padding and will begin with a two-byte instruction.
1165 Most of the functions in the Windows system DLLs in Windows XP SP2 or
1166 higher were compiled in this fashion.</dd>
Charles Davis970bfcc2010-10-25 15:37:09 +0000[diff] [blame]1167
Dan Gohman129bd562011-06-16 16:03:13 +0000[diff] [blame]1168 <dt><tt><b>nonlazybind</b></tt></dt>
1169 <dd>This attribute suppresses lazy symbol binding for the function. This
1170 may make calls to the function faster, at the cost of extra program
1171 startup time if the function is not called during program startup.</dd>
1172
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +0000[diff] [blame]1173 <dt><tt><b>inlinehint</b></tt></dt>
1174 <dd>This attribute indicates that the source code contained a hint that inlining
1175 this function is desirable (such as the "inline" keyword in C/C++). It
1176 is just a hint; it imposes no requirements on the inliner.</dd>
1177
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1178 <dt><tt><b>naked</b></tt></dt>
1179 <dd>This attribute disables prologue / epilogue emission for the function.
1180 This can have very system-specific consequences.</dd>
1181
1182 <dt><tt><b>noimplicitfloat</b></tt></dt>
1183 <dd>This attributes disables implicit floating point instructions.</dd>
1184
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1185 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1186 <dd>This attribute indicates that the inliner should never inline this
1187 function in any situation. This attribute may not be used together with
1188 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1189
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1190 <dt><tt><b>noredzone</b></tt></dt>
1191 <dd>This attribute indicates that the code generator should not use a red
1192 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1193
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1194 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1195 <dd>This function attribute indicates that the function never returns
1196 normally. This produces undefined behavior at runtime if the function
1197 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +0000[diff] [blame]1198
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1199 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1200 <dd>This function attribute indicates that the function never returns with an
1201 unwind or exceptional control flow. If the function does unwind, its
1202 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +0000[diff] [blame]1203
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1204 <dt><tt><b>optsize</b></tt></dt>
1205 <dd>This attribute suggests that optimization passes and code generator passes
1206 make choices that keep the code size of this function low, and otherwise
1207 do optimizations specifically to reduce code size.</dd>
1208
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1209 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1210 <dd>This attribute indicates that the function computes its result (or decides
1211 to unwind an exception) based strictly on its arguments, without
1212 dereferencing any pointer arguments or otherwise accessing any mutable
1213 state (e.g. memory, control registers, etc) visible to caller functions.
1214 It does not write through any pointer arguments
1215 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1216 changes any state visible to callers. This means that it cannot unwind
1217 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1218 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +0000[diff] [blame]1219
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1220 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1221 <dd>This attribute indicates that the function does not write through any
1222 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1223 arguments) or otherwise modify any state (e.g. memory, control registers,
1224 etc) visible to caller functions. It may dereference pointer arguments
1225 and read state that may be set in the caller. A readonly function always
1226 returns the same value (or unwinds an exception identically) when called
1227 with the same set of arguments and global state. It cannot unwind an
1228 exception by calling the <tt>C++</tt> exception throwing methods, but may
1229 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +0000[diff] [blame]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>
1258
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1259</dl>
1260
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1261</div>
1262
1263<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1264<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]1265 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1266</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1267
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1268<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1269
1270<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1271 the GCC "file scope inline asm" blocks. These blocks are internally
1272 concatenated by LLVM and treated as a single unit, but may be separated in
1273 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1274
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1275<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1276module asm "inline asm code goes here"
1277module asm "more can go here"
1278</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1279
1280<p>The strings can contain any character by escaping non-printable characters.
1281 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]1282 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1283
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1284<p>The inline asm code is simply printed to the machine code .s file when
1285 assembly code is generated.</p>
1286
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1287</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1288
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1289<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1290<h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1291 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1292</h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1293
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1294<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1295
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1296<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1297 data is to be laid out in memory. The syntax for the data layout is
1298 simply:</p>
1299
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1300<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1301target datalayout = "<i>layout specification</i>"
1302</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1303
1304<p>The <i>layout specification</i> consists of a list of specifications
1305 separated by the minus sign character ('-'). Each specification starts with
1306 a letter and may include other information after the letter to define some
1307 aspect of the data layout. The specifications accepted are as follows:</p>
1308
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1309<dl>
1310 <dt><tt>E</tt></dt>
1311 <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]1312 bits with the most significance have the lowest address location.</dd>
1313
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1314 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1315 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1316 the bits with the least significance have the lowest address
1317 location.</dd>
1318
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1319 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1320 <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]1321 <i>preferred</i> alignments. All sizes are in bits. Specifying
1322 the <i>pref</i> alignment is optional. If omitted, the
1323 preceding <tt>:</tt> should be omitted too.</dd>
1324
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1325 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1326 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1327 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1328
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1329 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1330 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1331 <i>size</i>.</dd>
1332
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1333 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1334 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +0000[diff] [blame]1335 <i>size</i>. Only values of <i>size</i> that are supported by the target
1336 will work. 32 (float) and 64 (double) are supported on all targets;
1337 80 or 128 (different flavors of long double) are also supported on some
1338 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1339
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1340 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1341 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1342 <i>size</i>.</dd>
1343
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1344 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1345 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1346 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1347
1348 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1349 <dd>This specifies a set of native integer widths for the target CPU
1350 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1351 "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]1352 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1353 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1354</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1355
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1356<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]1357 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1358 specifications in the <tt>datalayout</tt> keyword. The default specifications
1359 are given in this list:</p>
1360
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1361<ul>
1362 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +0000[diff] [blame]1363 <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]1364 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1365 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1366 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1367 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1368 <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]1369 alignment of 64-bits</li>
1370 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1371 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1372 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1373 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1374 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1375 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1376</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1377
1378<p>When LLVM is determining the alignment for a given type, it uses the
1379 following rules:</p>
1380
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1381<ol>
1382 <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]1383 specification is used.</li>
1384
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1385 <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]1386 smallest integer type that is larger than the bitwidth of the sought type
1387 is used. If none of the specifications are larger than the bitwidth then
1388 the the largest integer type is used. For example, given the default
1389 specifications above, the i7 type will use the alignment of i8 (next
1390 largest) while both i65 and i256 will use the alignment of i64 (largest
1391 specified).</li>
1392
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1393 <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]1394 largest vector type that is smaller than the sought vector type will be
1395 used as a fall back. This happens because &lt;128 x double&gt; can be
1396 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1397</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1398
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1399</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1400
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1401<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1402<h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1403 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1404</h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1405
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1406<div>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1407
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1408<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +0000[diff] [blame]1409with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1410is undefined. Pointer values are associated with address ranges
1411according to the following rules:</p>
1412
1413<ul>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1414 <li>A pointer value is associated with the addresses associated with
1415 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1416 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1417 range of the variable's storage.</li>
1418 <li>The result value of an allocation instruction is associated with
1419 the address range of the allocated storage.</li>
1420 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1421 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1422 <li>An integer constant other than zero or a pointer value returned
1423 from a function not defined within LLVM may be associated with address
1424 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1425 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1426 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1427</ul>
1428
1429<p>A pointer value is <i>based</i> on another pointer value according
1430 to the following rules:</p>
1431
1432<ul>
1433 <li>A pointer value formed from a
1434 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1435 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1436 <li>The result value of a
1437 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1438 of the <tt>bitcast</tt>.</li>
1439 <li>A pointer value formed by an
1440 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1441 pointer values that contribute (directly or indirectly) to the
1442 computation of the pointer's value.</li>
1443 <li>The "<i>based</i> on" relationship is transitive.</li>
1444</ul>
1445
1446<p>Note that this definition of <i>"based"</i> is intentionally
1447 similar to the definition of <i>"based"</i> in C99, though it is
1448 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1449
1450<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1451<tt><a href="#i_load">load</a></tt> merely indicates the size and
1452alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +0000[diff] [blame]1453interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1454<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1455and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1456
1457<p>Consequently, type-based alias analysis, aka TBAA, aka
1458<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1459LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1460additional information which specialized optimization passes may use
1461to implement type-based alias analysis.</p>
1462
1463</div>
1464
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1465<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1466<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1467 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1468</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1469
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1470<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1471
1472<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1473href="#i_store"><tt>store</tt></a>s, and <a
1474href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1475The optimizers must not change the number of volatile operations or change their
1476order of execution relative to other volatile operations. The optimizers
1477<i>may</i> change the order of volatile operations relative to non-volatile
1478operations. This is not Java's "volatile" and has no cross-thread
1479synchronization behavior.</p>
1480
1481</div>
1482
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1483<!-- ======================================================================= -->
1484<h3>
1485 <a name="memmodel">Memory Model for Concurrent Operations</a>
1486</h3>
1487
1488<div>
1489
1490<p>The LLVM IR does not define any way to start parallel threads of execution
1491or to register signal handlers. Nonetheless, there are platform-specific
1492ways to create them, and we define LLVM IR's behavior in their presence. This
1493model is inspired by the C++0x memory model.</p>
1494
1495<p>We define a <i>happens-before</i> partial order as the least partial order
1496that</p>
1497<ul>
1498 <li>Is a superset of single-thread program order, and</li>
1499 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1500 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1501 by platform-specific techniques, like pthread locks, thread
1502 creation, thread joining, etc., and by the atomic operations described
1503 in the <a href="#int_atomics">Atomic intrinsics</a> section.</li>
1504</ul>
1505
1506<p>Note that program order does not introduce <i>happens-before</i> edges
1507between a thread and signals executing inside that thread.</p>
1508
1509<p>Every (defined) read operation (load instructions, memcpy, atomic
1510loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1511(defined) write operations (store instructions, atomic
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1512stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1513initialized globals are considered to have a write of the initializer which is
1514atomic and happens before any other read or write of the memory in question.
1515For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1516any write to the same byte, except:</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1517
1518<ul>
1519 <li>If <var>write<sub>1</sub></var> happens before
1520 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1521 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1522 does not see <var>write<sub>1</sub></var>.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1523 <li>If <var>R<sub>byte</sub></var> happens before <var>write<sub>3</var>,
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1524 then <var>R<sub>byte</sub></var> does not see
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1525 <var>write<sub>3</sub></var>.
1526</ul>
1527
1528<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1529<ul>
1530 <li>If there is no write to the same byte that happens before
1531 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1532 <tt>undef</tt> for that byte.
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1533 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1534 <var>R<sub>byte</sub></var> returns the value written by that
1535 write.</li>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1536 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1537 <var>R<sub>byte</sub></var> may see are atomic, it chooses one of the
1538 values written. See the <a href="#int_atomics">Atomic intrinsics</a>
1539 section for additional guarantees on how the choice is made.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1540 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1541</ul>
1542
1543<p><var>R</var> returns the value composed of the series of bytes it read.
1544This implies that some bytes within the value may be <tt>undef</tt>
1545<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1546defines the semantics of the operation; it doesn't mean that targets will
1547emit more than one instruction to read the series of bytes.</p>
1548
1549<p>Note that in cases where none of the atomic intrinsics are used, this model
1550places only one restriction on IR transformations on top of what is required
1551for single-threaded execution: introducing a store to a byte which might not
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1552otherwise be stored to can introduce undefined behavior. (Specifically, in
1553the case where another thread might write to and read from an address,
1554introducing a store can change a load that may see exactly one write into
1555a load that may see multiple writes.)</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1556
1557<!-- FIXME: This model assumes all targets where concurrency is relevant have
1558a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1559none of the backends currently in the tree fall into this category; however,
1560there might be targets which care. If there are, we want a paragraph
1561like the following:
1562
1563Targets may specify that stores narrower than a certain width are not
1564available; on such a target, for the purposes of this model, treat any
1565non-atomic write with an alignment or width less than the minimum width
1566as if it writes to the relevant surrounding bytes.
1567-->
1568
1569</div>
1570
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1571</div>
1572
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1573<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1574<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1575<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1576
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1577<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1578
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1579<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1580 intermediate representation. Being typed enables a number of optimizations
1581 to be performed on the intermediate representation directly, without having
1582 to do extra analyses on the side before the transformation. A strong type
1583 system makes it easier to read the generated code and enables novel analyses
1584 and transformations that are not feasible to perform on normal three address
1585 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1586
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1587<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1588<h3>
1589 <a name="t_classifications">Type Classifications</a>
1590</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1591
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1592<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1593
1594<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1595
1596<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1597 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1598 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1599 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1600 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1601 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1602 </tr>
1603 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1604 <td><a href="#t_floating">floating point</a></td>
1605 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1606 </tr>
1607 <tr>
1608 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1609 <td><a href="#t_integer">integer</a>,
1610 <a href="#t_floating">floating point</a>,
1611 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +0000[diff] [blame]1612 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]1613 <a href="#t_struct">structure</a>,
1614 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1615 <a href="#t_label">label</a>,
1616 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1617 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1618 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1619 <tr>
1620 <td><a href="#t_primitive">primitive</a></td>
1621 <td><a href="#t_label">label</a>,
1622 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +0000[diff] [blame]1623 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1624 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1625 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1626 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1627 </tr>
1628 <tr>
1629 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1630 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1631 <a href="#t_function">function</a>,
1632 <a href="#t_pointer">pointer</a>,
1633 <a href="#t_struct">structure</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1634 <a href="#t_vector">vector</a>,
1635 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +0000[diff] [blame]1636 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1637 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1638 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1639</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1640
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1641<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1642 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1643 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1644
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1645</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1646
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1647<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1648<h3>
1649 <a name="t_primitive">Primitive Types</a>
1650</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1651
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1652<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1653
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1654<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1655 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1656
1657<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1658<h4>
1659 <a name="t_integer">Integer Type</a>
1660</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1661
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1662<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1663
1664<h5>Overview:</h5>
1665<p>The integer type is a very simple type that simply specifies an arbitrary
1666 bit width for the integer type desired. Any bit width from 1 bit to
1667 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1668
1669<h5>Syntax:</h5>
1670<pre>
1671 iN
1672</pre>
1673
1674<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1675 value.</p>
1676
1677<h5>Examples:</h5>
1678<table class="layout">
1679 <tr class="layout">
1680 <td class="left"><tt>i1</tt></td>
1681 <td class="left">a single-bit integer.</td>
1682 </tr>
1683 <tr class="layout">
1684 <td class="left"><tt>i32</tt></td>
1685 <td class="left">a 32-bit integer.</td>
1686 </tr>
1687 <tr class="layout">
1688 <td class="left"><tt>i1942652</tt></td>
1689 <td class="left">a really big integer of over 1 million bits.</td>
1690 </tr>
1691</table>
1692
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1693</div>
1694
1695<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1696<h4>
1697 <a name="t_floating">Floating Point Types</a>
1698</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1699
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1700<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1701
1702<table>
1703 <tbody>
1704 <tr><th>Type</th><th>Description</th></tr>
1705 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1706 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1707 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1708 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1709 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1710 </tbody>
1711</table>
1712
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1713</div>
1714
1715<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1716<h4>
1717 <a name="t_x86mmx">X86mmx Type</a>
1718</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1719
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1720<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1721
1722<h5>Overview:</h5>
1723<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>
1724
1725<h5>Syntax:</h5>
1726<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +0000[diff] [blame]1727 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1728</pre>
1729
1730</div>
1731
1732<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1733<h4>
1734 <a name="t_void">Void Type</a>
1735</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1736
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1737<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1738
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1739<h5>Overview:</h5>
1740<p>The void type does not represent any value and has no size.</p>
1741
1742<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1743<pre>
1744 void
1745</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1746
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1747</div>
1748
1749<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1750<h4>
1751 <a name="t_label">Label Type</a>
1752</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1753
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1754<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1755
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1756<h5>Overview:</h5>
1757<p>The label type represents code labels.</p>
1758
1759<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1760<pre>
1761 label
1762</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1763
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1764</div>
1765
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1766<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1767<h4>
1768 <a name="t_metadata">Metadata Type</a>
1769</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1770
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1771<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1772
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1773<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1774<p>The metadata type represents embedded metadata. No derived types may be
1775 created from metadata except for <a href="#t_function">function</a>
1776 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1777
1778<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1779<pre>
1780 metadata
1781</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1782
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1783</div>
1784
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1785</div>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1786
1787<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1788<h3>
1789 <a name="t_derived">Derived Types</a>
1790</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1791
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1792<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1793
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1794<p>The real power in LLVM comes from the derived types in the system. This is
1795 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1796 useful types. Each of these types contain one or more element types which
1797 may be a primitive type, or another derived type. For example, it is
1798 possible to have a two dimensional array, using an array as the element type
1799 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1800
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1801</div>
1802
1803
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1804<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1805<h4>
1806 <a name="t_aggregate">Aggregate Types</a>
1807</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1808
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1809<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1810
1811<p>Aggregate Types are a subset of derived types that can contain multiple
1812 member types. <a href="#t_array">Arrays</a>,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]1813 <a href="#t_struct">structs</a>, and <a href="#t_vector">vectors</a> are
1814 aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1815
1816</div>
1817
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1818<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1819<h4>
1820 <a name="t_array">Array Type</a>
1821</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1822
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1823<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1824
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1825<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1826<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1827 sequentially in memory. The array type requires a size (number of elements)
1828 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1829
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1830<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1831<pre>
1832 [&lt;# elements&gt; x &lt;elementtype&gt;]
1833</pre>
1834
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1835<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1836 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1837
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1838<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1839<table class="layout">
1840 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1841 <td class="left"><tt>[40 x i32]</tt></td>
1842 <td class="left">Array of 40 32-bit integer values.</td>
1843 </tr>
1844 <tr class="layout">
1845 <td class="left"><tt>[41 x i32]</tt></td>
1846 <td class="left">Array of 41 32-bit integer values.</td>
1847 </tr>
1848 <tr class="layout">
1849 <td class="left"><tt>[4 x i8]</tt></td>
1850 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1851 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1852</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1853<p>Here are some examples of multidimensional arrays:</p>
1854<table class="layout">
1855 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1856 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1857 <td class="left">3x4 array of 32-bit integer values.</td>
1858 </tr>
1859 <tr class="layout">
1860 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1861 <td class="left">12x10 array of single precision floating point values.</td>
1862 </tr>
1863 <tr class="layout">
1864 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1865 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1866 </tr>
1867</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1868
Dan Gohman7657f6b2009-11-09 19:01:53 +0000[diff] [blame]1869<p>There is no restriction on indexing beyond the end of the array implied by
1870 a static type (though there are restrictions on indexing beyond the bounds
1871 of an allocated object in some cases). This means that single-dimension
1872 'variable sized array' addressing can be implemented in LLVM with a zero
1873 length array type. An implementation of 'pascal style arrays' in LLVM could
1874 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1875
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1876</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1877
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1878<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1879<h4>
1880 <a name="t_function">Function Type</a>
1881</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1882
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1883<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1884
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1885<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1886<p>The function type can be thought of as a function signature. It consists of
1887 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]1888 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]1889
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1890<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1891<pre>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]1892 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1893</pre>
1894
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1895<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]1896 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
1897 which indicates that the function takes a variable number of arguments.
1898 Variable argument functions can access their arguments with
1899 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]1900 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1901 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1902
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1903<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1904<table class="layout">
1905 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1906 <td class="left"><tt>i32 (i32)</tt></td>
1907 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1908 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1909 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]1910 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]1911 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1912 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]1913 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
1914 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1915 </td>
1916 </tr><tr class="layout">
1917 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1918 <td class="left">A vararg function that takes at least one
1919 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
1920 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1921 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1922 </td>
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]1923 </tr><tr class="layout">
1924 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]1925 <td class="left">A function taking an <tt>i32</tt>, returning a
1926 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]1927 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1928 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1929</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1930
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1931</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1932
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1933<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1934<h4>
1935 <a name="t_struct">Structure Type</a>
1936</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1937
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1938<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1939
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1940<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1941<p>The structure type is used to represent a collection of data members together
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1942 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]1943
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]1944<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
1945 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
1946 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
1947 Structures in registers are accessed using the
1948 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
1949 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1950
1951<p>Structures may optionally be "packed" structures, which indicate that the
1952 alignment of the struct is one byte, and that there is no padding between
1953 the elements. In non-packed structs, padding between field types is defined
1954 by the target data string to match the underlying processor.</p>
1955
1956<p>Structures can either be "anonymous" or "named". An anonymous structure is
1957 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) and a named types
1958 are always defined at the top level with a name. Anonmyous types are uniqued
1959 by their contents and can never be recursive since there is no way to write
1960 one. Named types can be recursive.
1961</p>
1962
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1963<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1964<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1965 %T1 = type { &lt;type list&gt; } <i>; Named normal struct type</i>
1966 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Named packed struct type</i>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1967</pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1968
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1969<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1970<table class="layout">
1971 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1972 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1973 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1974 </tr>
1975 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1976 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1977 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1978 second element is a <a href="#t_pointer">pointer</a> to a
1979 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1980 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1981 </tr>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1982 <tr class="layout">
1983 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
1984 <td class="left">A packed struct known to be 5 bytes in size.</td>
1985 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1986</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1987
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1988</div>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]1989
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1990<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1991<h4>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]1992 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1993</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1994
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1995<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1996
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1997<h5>Overview:</h5>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]1998<p>Opaque structure types are used to represent named structure types that do
1999 not have a body specified. This corresponds (for example) to the C notion of
2000 a forward declared structure.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2001
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2002<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2003<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2004 %X = type opaque
2005 %52 = type opaque
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2006</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2007
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2008<h5>Examples:</h5>
2009<table class="layout">
2010 <tr class="layout">
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2011 <td class="left"><tt>opaque</tt></td>
2012 <td class="left">An opaque type.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2013 </tr>
2014</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2015
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2016</div>
2017
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2018
2019
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2020<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2021<h4>
2022 <a name="t_pointer">Pointer Type</a>
2023</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2024
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2025<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2026
2027<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +0000[diff] [blame]2028<p>The pointer type is used to specify memory locations.
2029 Pointers are commonly used to reference objects in memory.</p>
2030
2031<p>Pointer types may have an optional address space attribute defining the
2032 numbered address space where the pointed-to object resides. The default
2033 address space is number zero. The semantics of non-zero address
2034 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2035
2036<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2037 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2038
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2039<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2040<pre>
2041 &lt;type&gt; *
2042</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2043
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2044<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2045<table class="layout">
2046 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2047 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2048 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2049 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2050 </tr>
2051 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]2052 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2053 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2054 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2055 <tt>i32</tt>.</td>
2056 </tr>
2057 <tr class="layout">
2058 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2059 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2060 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2061 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2062</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2063
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2064</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2065
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2066<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2067<h4>
2068 <a name="t_vector">Vector Type</a>
2069</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2070
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2071<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2072
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2073<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2074<p>A vector type is a simple derived type that represents a vector of elements.
2075 Vector types are used when multiple primitive data are operated in parallel
2076 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +0000[diff] [blame]2077 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2078 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2079
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2080<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2081<pre>
2082 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2083</pre>
2084
Chris Lattner7d2e7be2010-10-10 18:20:35 +0000[diff] [blame]2085<p>The number of elements is a constant integer value larger than 0; elementtype
2086 may be any integer or floating point type. Vectors of size zero are not
2087 allowed, and pointers are not allowed as the element type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2088
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2089<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2090<table class="layout">
2091 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2092 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2093 <td class="left">Vector of 4 32-bit integer values.</td>
2094 </tr>
2095 <tr class="layout">
2096 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2097 <td class="left">Vector of 8 32-bit floating-point values.</td>
2098 </tr>
2099 <tr class="layout">
2100 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2101 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2102 </tr>
2103</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2104
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2105</div>
2106
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2107<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2108<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2109<!-- *********************************************************************** -->
2110
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2111<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2112
2113<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2114 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2115
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2116<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2117<h3>
2118 <a name="simpleconstants">Simple Constants</a>
2119</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2120
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2121<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2122
2123<dl>
2124 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2125 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]2126 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2127
2128 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2129 <dd>Standard integers (such as '4') are constants of
2130 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2131 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2132
2133 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2134 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2135 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2136 notation (see below). The assembler requires the exact decimal value of a
2137 floating-point constant. For example, the assembler accepts 1.25 but
2138 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2139 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2140
2141 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2142 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2143 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2144</dl>
2145
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2146<p>The one non-intuitive notation for constants is the hexadecimal form of
2147 floating point constants. For example, the form '<tt>double
2148 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2149 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2150 constants are required (and the only time that they are generated by the
2151 disassembler) is when a floating point constant must be emitted but it cannot
2152 be represented as a decimal floating point number in a reasonable number of
2153 digits. For example, NaN's, infinities, and other special values are
2154 represented in their IEEE hexadecimal format so that assembly and disassembly
2155 do not cause any bits to change in the constants.</p>
2156
Dale Johannesenbd5e5a82009-02-11 22:14:51 +0000[diff] [blame]2157<p>When using the hexadecimal form, constants of types float and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2158 represented using the 16-digit form shown above (which matches the IEEE754
2159 representation for double); float values must, however, be exactly
2160 representable as IEE754 single precision. Hexadecimal format is always used
2161 for long double, and there are three forms of long double. The 80-bit format
2162 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2163 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2164 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2165 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2166 currently supported target uses this format. Long doubles will only work if
2167 they match the long double format on your target. All hexadecimal formats
2168 are big-endian (sign bit at the left).</p>
2169
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]2170<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2171</div>
2172
2173<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2174<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]2175<a name="aggregateconstants"></a> <!-- old anchor -->
2176<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2177</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2178
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2179<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2180
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2181<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2182 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2183
2184<dl>
2185 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2186 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2187 type definitions (a comma separated list of elements, surrounded by braces
2188 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2189 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2190 Structure constants must have <a href="#t_struct">structure type</a>, and
2191 the number and types of elements must match those specified by the
2192 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2193
2194 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2195 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2196 definitions (a comma separated list of elements, surrounded by square
2197 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2198 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2199 the number and types of elements must match those specified by the
2200 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2201
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2202 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2203 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2204 definitions (a comma separated list of elements, surrounded by
2205 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2206 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2207 have <a href="#t_vector">vector type</a>, and the number and types of
2208 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2209
2210 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2211 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]2212 value to zero of <em>any</em> type, including scalar and
2213 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2214 This is often used to avoid having to print large zero initializers
2215 (e.g. for large arrays) and is always exactly equivalent to using explicit
2216 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +0000[diff] [blame]2217
2218 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +0000[diff] [blame]2219 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2220 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2221 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2222 be interpreted as part of the instruction stream, metadata is a place to
2223 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2224</dl>
2225
2226</div>
2227
2228<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2229<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2230 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2231</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2232
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2233<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2234
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2235<p>The addresses of <a href="#globalvars">global variables</a>
2236 and <a href="#functionstructure">functions</a> are always implicitly valid
2237 (link-time) constants. These constants are explicitly referenced when
2238 the <a href="#identifiers">identifier for the global</a> is used and always
2239 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2240 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2241
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2242<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]2243@X = global i32 17
2244@Y = global i32 42
2245@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2246</pre>
2247
2248</div>
2249
2250<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2251<h3>
2252 <a name="undefvalues">Undefined Values</a>
2253</h3>
2254
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2255<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2256
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2257<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]2258 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2259 Undefined values may be of any type (other than '<tt>label</tt>'
2260 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2261
Chris Lattnerc608cb12009-09-11 01:49:31 +0000[diff] [blame]2262<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2263 program is well defined no matter what value is used. This gives the
2264 compiler more freedom to optimize. Here are some examples of (potentially
2265 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2266
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2267
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2268<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2269 %A = add %X, undef
2270 %B = sub %X, undef
2271 %C = xor %X, undef
2272Safe:
2273 %A = undef
2274 %B = undef
2275 %C = undef
2276</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2277
2278<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]2279 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]2280
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2281<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2282 %A = or %X, undef
2283 %B = and %X, undef
2284Safe:
2285 %A = -1
2286 %B = 0
2287Unsafe:
2288 %A = undef
2289 %B = undef
2290</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2291
2292<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2293 For example, if <tt>%X</tt> has a zero bit, then the output of the
2294 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2295 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2296 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2297 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2298 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2299 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2300 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2301
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2302<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2303 %A = select undef, %X, %Y
2304 %B = select undef, 42, %Y
2305 %C = select %X, %Y, undef
2306Safe:
2307 %A = %X (or %Y)
2308 %B = 42 (or %Y)
2309 %C = %Y
2310Unsafe:
2311 %A = undef
2312 %B = undef
2313 %C = undef
2314</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2315
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2316<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2317 branch) conditions can go <em>either way</em>, but they have to come from one
2318 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2319 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2320 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2321 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2322 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2323 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2324
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2325<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2326 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2327
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2328 %B = undef
2329 %C = xor %B, %B
2330
2331 %D = undef
2332 %E = icmp lt %D, 4
2333 %F = icmp gte %D, 4
2334
2335Safe:
2336 %A = undef
2337 %B = undef
2338 %C = undef
2339 %D = undef
2340 %E = undef
2341 %F = undef
2342</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2343
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2344<p>This example points out that two '<tt>undef</tt>' operands are not
2345 necessarily the same. This can be surprising to people (and also matches C
2346 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2347 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2348 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2349 its value over its "live range". This is true because the variable doesn't
2350 actually <em>have a live range</em>. Instead, the value is logically read
2351 from arbitrary registers that happen to be around when needed, so the value
2352 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2353 need to have the same semantics or the core LLVM "replace all uses with"
2354 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2355
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2356<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2357 %A = fdiv undef, %X
2358 %B = fdiv %X, undef
2359Safe:
2360 %A = undef
2361b: unreachable
2362</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2363
2364<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2365 value</em> and <em>undefined behavior</em>. An undefined value (like
2366 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2367 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2368 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2369 defined on SNaN's. However, in the second example, we can make a more
2370 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2371 arbitrary value, we are allowed to assume that it could be zero. Since a
2372 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2373 the operation does not execute at all. This allows us to delete the divide and
2374 all code after it. Because the undefined operation "can't happen", the
2375 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2376
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2377<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2378a: store undef -> %X
2379b: store %X -> undef
2380Safe:
2381a: &lt;deleted&gt;
2382b: unreachable
2383</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2384
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2385<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2386 undefined value can be assumed to not have any effect; we can assume that the
2387 value is overwritten with bits that happen to match what was already there.
2388 However, a store <em>to</em> an undefined location could clobber arbitrary
2389 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2390
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2391</div>
2392
2393<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2394<h3>
2395 <a name="trapvalues">Trap Values</a>
2396</h3>
2397
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2398<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2399
Dan Gohmanc68ce062010-04-26 20:21:21 +0000[diff] [blame]2400<p>Trap values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2401 instead of representing an unspecified bit pattern, they represent the
2402 fact that an instruction or constant expression which cannot evoke side
2403 effects has nevertheless detected a condition which results in undefined
Dan Gohmanc68ce062010-04-26 20:21:21 +0000[diff] [blame]2404 behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2405
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2406<p>There is currently no way of representing a trap value in the IR; they
Dan Gohman855abed2010-05-03 14:51:43 +0000[diff] [blame]2407 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2408 <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]2409
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2410<p>Trap value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2411
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2412<ul>
2413<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2414 their operands.</li>
2415
2416<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2417 to their dynamic predecessor basic block.</li>
2418
2419<li>Function arguments depend on the corresponding actual argument values in
2420 the dynamic callers of their functions.</li>
2421
2422<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2423 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2424 control back to them.</li>
2425
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2426<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2427 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2428 or exception-throwing call instructions that dynamically transfer control
2429 back to them.</li>
2430
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2431<li>Non-volatile loads and stores depend on the most recent stores to all of the
2432 referenced memory addresses, following the order in the IR
2433 (including loads and stores implied by intrinsics such as
2434 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2435
Dan Gohman7c24ff12010-05-03 14:59:34 +0000[diff] [blame]2436<!-- TODO: In the case of multiple threads, this only applies if the store
2437 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2438
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2439<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2440
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2441<li>An instruction with externally visible side effects depends on the most
2442 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]2443 the order in the IR. (This includes
2444 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2445
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2446<li>An instruction <i>control-depends</i> on a
2447 <a href="#terminators">terminator instruction</a>
2448 if the terminator instruction has multiple successors and the instruction
2449 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2450 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2451
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2452<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2453 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2454 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2455 successor.</li>
2456
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2457<li>Dependence is transitive.</li>
2458
2459</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2460
2461<p>Whenever a trap value is generated, all values which depend on it evaluate
2462 to trap. If they have side effects, the evoke their side effects as if each
2463 operand with a trap value were undef. If they have externally-visible side
2464 effects, the behavior is undefined.</p>
2465
2466<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +0000[diff] [blame]2467
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2468<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2469entry:
2470 %trap = sub nuw i32 0, 1 ; Results in a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2471 %still_trap = and i32 %trap, 0 ; Whereas (and i32 undef, 0) would return 0.
2472 %trap_yet_again = getelementptr i32* @h, i32 %still_trap
2473 store i32 0, i32* %trap_yet_again ; undefined behavior
2474
2475 store i32 %trap, i32* @g ; Trap value conceptually stored to memory.
2476 %trap2 = load i32* @g ; Returns a trap value, not just undef.
2477
2478 volatile store i32 %trap, i32* @g ; External observation; undefined behavior.
2479
2480 %narrowaddr = bitcast i32* @g to i16*
2481 %wideaddr = bitcast i32* @g to i64*
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2482 %trap3 = load i16* %narrowaddr ; Returns a trap value.
2483 %trap4 = load i64* %wideaddr ; Returns a trap value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2484
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2485 %cmp = icmp slt i32 %trap, 0 ; Returns a trap value.
2486 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2487
2488true:
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2489 volatile store i32 0, i32* @g ; This is control-dependent on %cmp, so
2490 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2491 br label %end
2492
2493end:
2494 %p = phi i32 [ 0, %entry ], [ 1, %true ]
2495 ; Both edges into this PHI are
2496 ; control-dependent on %cmp, so this
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2497 ; always results in a trap value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2498
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2499 volatile store i32 0, i32* @g ; This would depend on the store in %true
2500 ; if %cmp is true, or the store in %entry
2501 ; otherwise, so this is undefined behavior.
2502
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2503 br i1 %cmp, label %second_true, label %second_end
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2504 ; The same branch again, but this time the
2505 ; true block doesn't have side effects.
2506
2507second_true:
2508 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2509 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2510
2511second_end:
2512 volatile store i32 0, i32* @g ; This time, the instruction always depends
2513 ; on the store in %end. Also, it is
2514 ; control-equivalent to %end, so this is
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2515 ; well-defined (again, ignoring earlier
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2516 ; undefined behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2517</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2518
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2519</div>
2520
2521<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2522<h3>
2523 <a name="blockaddress">Addresses of Basic Blocks</a>
2524</h3>
2525
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2526<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2527
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2528<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2529
2530<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +0000[diff] [blame]2531 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2532 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2533
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2534<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2535 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2536 comparisons against null. Pointer equality tests between labels addresses
2537 results in undefined behavior &mdash; though, again, comparison against null
2538 is ok, and no label is equal to the null pointer. This may be passed around
2539 as an opaque pointer sized value as long as the bits are not inspected. This
2540 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2541 long as the original value is reconstituted before the <tt>indirectbr</tt>
2542 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2543
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2544<p>Finally, some targets may provide defined semantics when using the value as
2545 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2546
2547</div>
2548
2549
2550<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2551<h3>
2552 <a name="constantexprs">Constant Expressions</a>
2553</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2554
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2555<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2556
2557<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2558 to be used as constants. Constant expressions may be of
2559 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2560 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2561 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2562
2563<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2564 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2565 <dd>Truncate a constant to another type. The bit size of CST must be larger
2566 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2567
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2568 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2569 <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]2570 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2571
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2572 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2573 <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]2574 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2575
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2576 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2577 <dd>Truncate a floating point constant to another floating point type. The
2578 size of CST must be larger than the size of TYPE. Both types must be
2579 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2580
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2581 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2582 <dd>Floating point extend a constant to another type. The size of CST must be
2583 smaller or equal to the size of TYPE. Both types must be floating
2584 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2585
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2586 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2587 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2588 constant. TYPE must be a scalar or vector integer type. CST must be of
2589 scalar or vector floating point type. Both CST and TYPE must be scalars,
2590 or vectors of the same number of elements. If the value won't fit in the
2591 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2592
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2593 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2594 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2595 constant. TYPE must be a scalar or vector integer type. CST must be of
2596 scalar or vector floating point type. Both CST and TYPE must be scalars,
2597 or vectors of the same number of elements. If the value won't fit in the
2598 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2599
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2600 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2601 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2602 constant. TYPE must be a scalar or vector floating point type. CST must be
2603 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2604 vectors of the same number of elements. If the value won't fit in the
2605 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2606
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2607 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2608 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2609 constant. TYPE must be a scalar or vector floating point type. CST must be
2610 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2611 vectors of the same number of elements. If the value won't fit in the
2612 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2613
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2614 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2615 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2616 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2617 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2618 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2619
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2620 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2621 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2622 type. CST must be of integer type. The CST value is zero extended,
2623 truncated, or unchanged to make it fit in a pointer size. This one is
2624 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2625
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2626 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +0000[diff] [blame]2627 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2628 are the same as those for the <a href="#i_bitcast">bitcast
2629 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2630
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2631 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2632 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2633 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2634 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2635 instruction, the index list may have zero or more indexes, which are
2636 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2637
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2638 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2639 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2640
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2641 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2642 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2643
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2644 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2645 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2646
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2647 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2648 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2649 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2650
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2651 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2652 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2653 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2654
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2655 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2656 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2657 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2658
Nick Lewycky9e130ce2010-05-29 06:44:15 +0000[diff] [blame]2659 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2660 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2661 constants. The index list is interpreted in a similar manner as indices in
2662 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2663 index value must be specified.</dd>
2664
2665 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2666 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2667 constants. The index list is interpreted in a similar manner as indices in
2668 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2669 index value must be specified.</dd>
2670
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2671 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2672 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2673 be any of the <a href="#binaryops">binary</a>
2674 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2675 on operands are the same as those for the corresponding instruction
2676 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2677</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2678
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2679</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]2680
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2681</div>
2682
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2683<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2684<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2685<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2686<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2687<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2688<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2689<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2690</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2691
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2692<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2693
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2694<p>LLVM supports inline assembler expressions (as opposed
2695 to <a href="#moduleasm"> Module-Level Inline Assembly</a>) through the use of
2696 a special value. This value represents the inline assembler as a string
2697 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2698 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2699 expression has side effects, and a flag indicating whether the function
2700 containing the asm needs to align its stack conservatively. An example
2701 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2702
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2703<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2704i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2705</pre>
2706
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2707<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2708 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2709 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2710
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2711<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2712%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]2713</pre>
2714
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2715<p>Inline asms with side effects not visible in the constraint list must be
2716 marked as having side effects. This is done through the use of the
2717 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2718
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2719<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2720call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2721</pre>
2722
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2723<p>In some cases inline asms will contain code that will not work unless the
2724 stack is aligned in some way, such as calls or SSE instructions on x86,
2725 yet will not contain code that does that alignment within the asm.
2726 The compiler should make conservative assumptions about what the asm might
2727 contain and should generate its usual stack alignment code in the prologue
2728 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2729
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2730<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2731call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2732</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2733
2734<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2735 first.</p>
2736
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2737<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2738 documented here. Constraints on what can be done (e.g. duplication, moving,
2739 etc need to be documented). This is probably best done by reference to
2740 another document that covers inline asm from a holistic perspective.</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2741
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2742<h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2743<a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2744</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2745
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2746<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2747
2748<p>The call instructions that wrap inline asm nodes may have a "!srcloc" MDNode
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2749 attached to it that contains a list of constant integers. If present, the
2750 code generator will use the integer as the location cookie value when report
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2751 errors through the LLVMContext error reporting mechanisms. This allows a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]2752 front-end to correlate backend errors that occur with inline asm back to the
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2753 source code that produced it. For example:</p>
2754
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2755<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2756call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2757...
2758!42 = !{ i32 1234567 }
2759</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2760
2761<p>It is up to the front-end to make sense of the magic numbers it places in the
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2762 IR. If the MDNode contains multiple constants, the code generator will use
2763 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]2764
2765</div>
2766
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2767</div>
2768
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2769<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2770<h3>
2771 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2772</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2773
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2774<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2775
2776<p>LLVM IR allows metadata to be attached to instructions in the program that
2777 can convey extra information about the code to the optimizers and code
2778 generator. One example application of metadata is source-level debug
2779 information. There are two metadata primitives: strings and nodes. All
2780 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2781 preceding exclamation point ('<tt>!</tt>').</p>
2782
2783<p>A metadata string is a string surrounded by double quotes. It can contain
2784 any character by escaping non-printable characters with "\xx" where "xx" is
2785 the two digit hex code. For example: "<tt>!"test\00"</tt>".</p>
2786
2787<p>Metadata nodes are represented with notation similar to structure constants
2788 (a comma separated list of elements, surrounded by braces and preceded by an
2789 exclamation point). For example: "<tt>!{ metadata !"test\00", i32
2790 10}</tt>". Metadata nodes can have any values as their operand.</p>
2791
2792<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2793 metadata nodes, which can be looked up in the module symbol table. For
2794 example: "<tt>!foo = metadata !{!4, !3}</tt>".
2795
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2796<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2797 function is using two metadata arguments.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2798
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2799<div class="doc_code">
2800<pre>
2801call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2802</pre>
2803</div>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2804
2805<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2806 attached with <tt>add</tt> instruction using <tt>!dbg</tt> identifier.</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2807
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2808<div class="doc_code">
2809<pre>
2810%indvar.next = add i64 %indvar, 1, !dbg !21
2811</pre>
2812</div>
2813
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2814</div>
2815
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2816</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2817
2818<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2819<h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2820 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2821</h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2822<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2823<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2824<p>LLVM has a number of "magic" global variables that contain data that affect
2825code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2826of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
2827section and all globals that start with "<tt>llvm.</tt>" are reserved for use
2828by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2829
2830<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2831<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2832<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2833</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2834
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2835<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2836
2837<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
2838href="#linkage_appending">appending linkage</a>. This array contains a list of
2839pointers to global variables and functions which may optionally have a pointer
2840cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
2841
2842<pre>
2843 @X = global i8 4
2844 @Y = global i32 123
2845
2846 @llvm.used = appending global [2 x i8*] [
2847 i8* @X,
2848 i8* bitcast (i32* @Y to i8*)
2849 ], section "llvm.metadata"
2850</pre>
2851
2852<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
2853compiler, assembler, and linker are required to treat the symbol as if there is
2854a reference to the global that it cannot see. For example, if a variable has
2855internal linkage and no references other than that from the <tt>@llvm.used</tt>
2856list, it cannot be deleted. This is commonly used to represent references from
2857inline asms and other things the compiler cannot "see", and corresponds to
2858"attribute((used))" in GNU C.</p>
2859
2860<p>On some targets, the code generator must emit a directive to the assembler or
2861object file to prevent the assembler and linker from molesting the symbol.</p>
2862
2863</div>
2864
2865<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2866<h3>
2867 <a name="intg_compiler_used">
2868 The '<tt>llvm.compiler.used</tt>' Global Variable
2869 </a>
2870</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2871
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2872<div>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]2873
2874<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
2875<tt>@llvm.used</tt> directive, except that it only prevents the compiler from
2876touching the symbol. On targets that support it, this allows an intelligent
2877linker to optimize references to the symbol without being impeded as it would be
2878by <tt>@llvm.used</tt>.</p>
2879
2880<p>This is a rare construct that should only be used in rare circumstances, and
2881should not be exposed to source languages.</p>
2882
2883</div>
2884
2885<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2886<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2887<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2888</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2889
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2890<div>
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2891<pre>
2892%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]2893@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2894</pre>
2895<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor functions and associated priorities. The functions referenced by this array will be called in ascending order of priority (i.e. lowest first) when the module is loaded. The order of functions with the same priority is not defined.
2896</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2897
2898</div>
2899
2900<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2901<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2902<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2903</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2904
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2905<div>
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2906<pre>
2907%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]2908@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2909</pre>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2910
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]2911<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions and associated priorities. The functions referenced by this array will be called in descending order of priority (i.e. highest first) when the module is loaded. The order of functions with the same priority is not defined.
2912</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2913
2914</div>
2915
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2916</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]2917
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2918<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2919<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2920<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2921
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2922<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2923
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2924<p>The LLVM instruction set consists of several different classifications of
2925 instructions: <a href="#terminators">terminator
2926 instructions</a>, <a href="#binaryops">binary instructions</a>,
2927 <a href="#bitwiseops">bitwise binary instructions</a>,
2928 <a href="#memoryops">memory instructions</a>, and
2929 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2930
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2931<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2932<h3>
2933 <a name="terminators">Terminator Instructions</a>
2934</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2935
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2936<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2937
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2938<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
2939 in a program ends with a "Terminator" instruction, which indicates which
2940 block should be executed after the current block is finished. These
2941 terminator instructions typically yield a '<tt>void</tt>' value: they produce
2942 control flow, not values (the one exception being the
2943 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
2944
Duncan Sands83821c82010-04-15 20:35:54 +0000[diff] [blame]2945<p>There are seven different terminator instructions: the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2946 '<a href="#i_ret"><tt>ret</tt></a>' instruction, the
2947 '<a href="#i_br"><tt>br</tt></a>' instruction, the
2948 '<a href="#i_switch"><tt>switch</tt></a>' instruction, the
Bill Wendling21c346e2009-11-02 00:25:26 +0000[diff] [blame]2949 '<a href="#i_indirectbr">'<tt>indirectbr</tt></a>' Instruction, the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2950 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the
2951 '<a href="#i_unwind"><tt>unwind</tt></a>' instruction, and the
2952 '<a href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2953
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2954<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2955<h4>
2956 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
2957</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2958
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2959<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2960
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2961<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]2962<pre>
2963 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]2964 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2965</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2966
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2967<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2968<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
2969 a value) from a function back to the caller.</p>
2970
2971<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
2972 value and then causes control flow, and one that just causes control flow to
2973 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2974
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2975<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2976<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
2977 return value. The type of the return value must be a
2978 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]2979
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2980<p>A function is not <a href="#wellformed">well formed</a> if it it has a
2981 non-void return type and contains a '<tt>ret</tt>' instruction with no return
2982 value or a return value with a type that does not match its type, or if it
2983 has a void return type and contains a '<tt>ret</tt>' instruction with a
2984 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2985
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2986<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2987<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
2988 the calling function's context. If the caller is a
2989 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
2990 instruction after the call. If the caller was an
2991 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
2992 the beginning of the "normal" destination block. If the instruction returns
2993 a value, that value shall set the call or invoke instruction's return
2994 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2995
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2996<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2997<pre>
2998 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2999 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +0000[diff] [blame]3000 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]3001</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]3002
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3003</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3004<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3005<h4>
3006 <a name="i_br">'<tt>br</tt>' Instruction</a>
3007</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3008
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3009<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3010
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3011<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3012<pre>
Bill Wendlingb3aa4712011-07-26 10:41:15 +0000[diff] [blame]3013 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3014 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3015</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3016
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3017<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3018<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3019 different basic block in the current function. There are two forms of this
3020 instruction, corresponding to a conditional branch and an unconditional
3021 branch.</p>
3022
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3023<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3024<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3025 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3026 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3027 target.</p>
3028
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3029<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3030<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3031 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3032 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3033 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3034
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3035<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]3036<pre>
3037Test:
3038 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3039 br i1 %cond, label %IfEqual, label %IfUnequal
3040IfEqual:
3041 <a href="#i_ret">ret</a> i32 1
3042IfUnequal:
3043 <a href="#i_ret">ret</a> i32 0
3044</pre>
3045
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3046</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3047
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3048<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3049<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3050 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3051</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3052
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3053<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3054
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3055<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3056<pre>
3057 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3058</pre>
3059
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3060<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3061<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]3062 several different places. It is a generalization of the '<tt>br</tt>'
3063 instruction, allowing a branch to occur to one of many possible
3064 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3065
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3066<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3067<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3068 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3069 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3070 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3071
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3072<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3073<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3074 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3075 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3076 transferred to the corresponding destination; otherwise, control flow is
3077 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3078
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3079<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3080<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3081 <tt>switch</tt> instruction, this instruction may be code generated in
3082 different ways. For example, it could be generated as a series of chained
3083 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3084
3085<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3086<pre>
3087 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3088 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3089 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3090
3091 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3092 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3093
3094 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3095 switch i32 %val, label %otherwise [ i32 0, label %onzero
3096 i32 1, label %onone
3097 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3098</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3099
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3100</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3101
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3102
3103<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3104<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3105 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3106</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3107
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3108<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3109
3110<h5>Syntax:</h5>
3111<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3112 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]3113</pre>
3114
3115<h5>Overview:</h5>
3116
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3117<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3118 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]3119 "<tt>address</tt>". Address must be derived from a <a
3120 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3121
3122<h5>Arguments:</h5>
3123
3124<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3125 rest of the arguments indicate the full set of possible destinations that the
3126 address may point to. Blocks are allowed to occur multiple times in the
3127 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3128
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3129<p>This destination list is required so that dataflow analysis has an accurate
3130 understanding of the CFG.</p>
3131
3132<h5>Semantics:</h5>
3133
3134<p>Control transfers to the block specified in the address argument. All
3135 possible destination blocks must be listed in the label list, otherwise this
3136 instruction has undefined behavior. This implies that jumps to labels
3137 defined in other functions have undefined behavior as well.</p>
3138
3139<h5>Implementation:</h5>
3140
3141<p>This is typically implemented with a jump through a register.</p>
3142
3143<h5>Example:</h5>
3144<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3145 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3146</pre>
3147
3148</div>
3149
3150
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3151<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3152<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3153 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3154</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3155
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3156<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3157
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3158<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3159<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3160 &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]3161 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3162</pre>
3163
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3164<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3165<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3166 function, with the possibility of control flow transfer to either the
3167 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3168 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3169 control flow will return to the "normal" label. If the callee (or any
3170 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3171 instruction, control is interrupted and continued at the dynamically nearest
3172 "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3173
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3174<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3175<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3176
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3177<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3178 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3179 convention</a> the call should use. If none is specified, the call
3180 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3181
3182 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3183 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3184 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3185
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3186 <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]3187 function value being invoked. In most cases, this is a direct function
3188 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3189 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3190
3191 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3192 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3193
3194 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]3195 signature argument types and parameter attributes. All arguments must be
3196 of <a href="#t_firstclass">first class</a> type. If the function
3197 signature indicates the function accepts a variable number of arguments,
3198 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3199
3200 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3201 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3202
3203 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3204 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3205
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3206 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3207 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3208 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3209</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3210
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3211<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3212<p>This instruction is designed to operate as a standard
3213 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3214 primary difference is that it establishes an association with a label, which
3215 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3216
3217<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3218 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3219 exception. Additionally, this is important for implementation of
3220 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3221
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3222<p>For the purposes of the SSA form, the definition of the value returned by the
3223 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3224 block to the "normal" label. If the callee unwinds then no return value is
3225 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +0000[diff] [blame]3226
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3227<p>Note that the code generator does not yet completely support unwind, and
3228that the invoke/unwind semantics are likely to change in future versions.</p>
3229
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3230<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3231<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3232 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3233 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3234 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3235 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3236</pre>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3237
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3238</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3239
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3240<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3241
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3242<h4>
3243 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3244</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3245
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3246<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3247
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3248<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3249<pre>
3250 unwind
3251</pre>
3252
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3253<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3254<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3255 at the first callee in the dynamic call stack which used
3256 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3257 This is primarily used to implement exception handling.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3258
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3259<h5>Semantics:</h5>
Chris Lattner72ed2002008-04-19 21:01:16 +0000[diff] [blame]3260<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3261 immediately halt. The dynamic call stack is then searched for the
3262 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3263 Once found, execution continues at the "exceptional" destination block
3264 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3265 instruction in the dynamic call chain, undefined behavior results.</p>
3266
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3267<p>Note that the code generator does not yet completely support unwind, and
3268that the invoke/unwind semantics are likely to change in future versions.</p>
3269
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3270</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3271
3272<!-- _______________________________________________________________________ -->
3273
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3274<h4>
3275 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3276</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3277
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3278<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3279
3280<h5>Syntax:</h5>
3281<pre>
3282 unreachable
3283</pre>
3284
3285<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3286<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3287 instruction is used to inform the optimizer that a particular portion of the
3288 code is not reachable. This can be used to indicate that the code after a
3289 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3290
3291<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3292<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3293
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3294</div>
3295
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3296</div>
3297
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3298<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3299<h3>
3300 <a name="binaryops">Binary Operations</a>
3301</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3302
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3303<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3304
3305<p>Binary operators are used to do most of the computation in a program. They
3306 require two operands of the same type, execute an operation on them, and
3307 produce a single value. The operands might represent multiple data, as is
3308 the case with the <a href="#t_vector">vector</a> data type. The result value
3309 has the same type as its operands.</p>
3310
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3311<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3312
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3313<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3314<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3315 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3316</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3317
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3318<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3319
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3320<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3321<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3322 &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]3323 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3324 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3325 &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]3326</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3327
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3328<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3329<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3330
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3331<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3332<p>The two arguments to the '<tt>add</tt>' instruction must
3333 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3334 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3335
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3336<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3337<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3338
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3339<p>If the sum has unsigned overflow, the result returned is the mathematical
3340 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]3341
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3342<p>Because LLVM integers use a two's complement representation, this instruction
3343 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3344
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3345<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3346 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3347 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3348 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3349 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3350
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3351<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3352<pre>
3353 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3354</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3355
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3356</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3357
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3358<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3359<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3360 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3361</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3362
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3363<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3364
3365<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3366<pre>
3367 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3368</pre>
3369
3370<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3371<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3372
3373<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3374<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3375 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3376 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3377
3378<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3379<p>The value produced is the floating point sum of the two operands.</p>
3380
3381<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3382<pre>
3383 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3384</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3385
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3386</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3387
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3388<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3389<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3390 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3391</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3392
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3393<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3394
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3395<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3396<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3397 &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]3398 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3399 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3400 &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]3401</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3402
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3403<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3404<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3405 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3406
3407<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3408 '<tt>neg</tt>' instruction present in most other intermediate
3409 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3410
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3411<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3412<p>The two arguments to the '<tt>sub</tt>' instruction must
3413 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3414 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3415
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3416<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3417<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3418
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3419<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3420 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3421 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3422
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3423<p>Because LLVM integers use a two's complement representation, this instruction
3424 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3425
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3426<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3427 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3428 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3429 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3430 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3431
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3432<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]3433<pre>
3434 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3435 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3436</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3437
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3438</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3439
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3440<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3441<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3442 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3443</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3444
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3445<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3446
3447<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3448<pre>
3449 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3450</pre>
3451
3452<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3453<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3454 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3455
3456<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3457 '<tt>fneg</tt>' instruction present in most other intermediate
3458 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3459
3460<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]3461<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3462 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3463 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3464
3465<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3466<p>The value produced is the floating point difference of the two operands.</p>
3467
3468<h5>Example:</h5>
3469<pre>
3470 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3471 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3472</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3473
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3474</div>
3475
3476<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3477<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3478 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3479</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3480
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3481<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3482
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3483<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3484<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3485 &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]3486 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3487 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3488 &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]3489</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3490
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3491<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3492<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3493
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3494<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3495<p>The two arguments to the '<tt>mul</tt>' instruction must
3496 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3497 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3498
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3499<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3500<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3501
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3502<p>If the result of the multiplication has unsigned overflow, the result
3503 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3504 width of the result.</p>
3505
3506<p>Because LLVM integers use a two's complement representation, and the result
3507 is the same width as the operands, this instruction returns the correct
3508 result for both signed and unsigned integers. If a full product
3509 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3510 be sign-extended or zero-extended as appropriate to the width of the full
3511 product.</p>
3512
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3513<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3514 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3515 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3516 is a <a href="#trapvalues">trap value</a> if unsigned and/or signed overflow,
3517 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3518
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3519<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3520<pre>
3521 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3522</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3523
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3524</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3525
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3526<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3527<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3528 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3529</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3530
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3531<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3532
3533<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3534<pre>
3535 &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]3536</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3537
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3538<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3539<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3540
3541<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3542<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3543 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3544 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3545
3546<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3547<p>The value produced is the floating point product of the two operands.</p>
3548
3549<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3550<pre>
3551 &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]3552</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3553
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3554</div>
3555
3556<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3557<h4>
3558 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3559</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3560
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3561<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3562
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3563<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3564<pre>
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3565 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3566 &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]3567</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3568
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3569<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3570<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3571
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3572<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3573<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3574 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3575 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3576
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3577<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3578<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3579
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3580<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3581 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3582
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3583<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3584
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3585<p>If the <tt>exact</tt> keyword is present, the result value of the
3586 <tt>udiv</tt> is a <a href="#trapvalues">trap value</a> if %op1 is not a
3587 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3588
3589
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3590<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3591<pre>
3592 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3593</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3594
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3595</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3596
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3597<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3598<h4>
3599 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3600</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3601
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3602<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3603
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3604<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3605<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3606 &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]3607 &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]3608</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3609
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3610<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3611<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3612
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3613<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3614<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3615 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3616 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3617
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3618<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3619<p>The value produced is the signed integer quotient of the two operands rounded
3620 towards zero.</p>
3621
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3622<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3623 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3624
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3625<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3626 undefined behavior; this is a rare case, but can occur, for example, by doing
3627 a 32-bit division of -2147483648 by -1.</p>
3628
Dan Gohman9c5beed2009-07-22 00:04:19 +0000[diff] [blame]3629<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]3630 <tt>sdiv</tt> is a <a href="#trapvalues">trap value</a> if the result would
Dan Gohman38da9272010-07-11 00:08:34 +0000[diff] [blame]3631 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3632
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3633<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3634<pre>
3635 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3636</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3637
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3638</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3639
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3640<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3641<h4>
3642 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3643</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3644
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3645<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3646
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3647<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3648<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3649 &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]3650</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3651
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3652<h5>Overview:</h5>
3653<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3654
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3655<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3656<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3657 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3658 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3659
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3660<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3661<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3662
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3663<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3664<pre>
3665 &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]3666</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3667
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3668</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3669
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3670<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3671<h4>
3672 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3673</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3674
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3675<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3676
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3677<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3678<pre>
3679 &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]3680</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3681
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3682<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3683<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3684 division of its two arguments.</p>
3685
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3686<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3687<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3688 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3689 values. Both arguments must have identical types.</p>
3690
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3691<h5>Semantics:</h5>
3692<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3693 This instruction always performs an unsigned division to get the
3694 remainder.</p>
3695
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3696<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3697 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3698
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3699<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3700
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3701<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3702<pre>
3703 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3704</pre>
3705
3706</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3707
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3708<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3709<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3710 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3711</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3712
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3713<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3714
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3715<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3716<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3717 &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]3718</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3719
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3720<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3721<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
3722 division of its two operands. This instruction can also take
3723 <a href="#t_vector">vector</a> versions of the values in which case the
3724 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]3725
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3726<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3727<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3728 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3729 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3730
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3731<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3732<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +0000[diff] [blame]3733 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
3734 <i>modulo</i> operator (where the result is either zero or has the same sign
3735 as the divisor, <tt>op2</tt>) of a value.
3736 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3737 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
3738 Math Forum</a>. For a table of how this is implemented in various languages,
3739 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
3740 Wikipedia: modulo operation</a>.</p>
3741
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3742<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3743 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
3744
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3745<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3746 Overflow also leads to undefined behavior; this is a rare case, but can
3747 occur, for example, by taking the remainder of a 32-bit division of
3748 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
3749 lets srem be implemented using instructions that return both the result of
3750 the division and the remainder.)</p>
3751
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3752<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3753<pre>
3754 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3755</pre>
3756
3757</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3758
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3759<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3760<h4>
3761 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
3762</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3763
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3764<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3765
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3766<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3767<pre>
3768 &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]3769</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3770
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3771<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3772<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
3773 its two operands.</p>
3774
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3775<h5>Arguments:</h5>
3776<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3777 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3778 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3779
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3780<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3781<p>This instruction returns the <i>remainder</i> of a division. The remainder
3782 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3783
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3784<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3785<pre>
3786 &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]3787</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3788
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3789</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]3790
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3791</div>
3792
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]3793<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3794<h3>
3795 <a name="bitwiseops">Bitwise Binary Operations</a>
3796</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3797
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3798<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3799
3800<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
3801 program. They are generally very efficient instructions and can commonly be
3802 strength reduced from other instructions. They require two operands of the
3803 same type, execute an operation on them, and produce a single value. The
3804 resulting value is the same type as its operands.</p>
3805
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3806<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3807<h4>
3808 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
3809</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3810
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3811<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3812
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3813<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3814<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3815 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3816 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3817 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3818 &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]3819</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3820
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3821<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3822<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
3823 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3824
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3825<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3826<p>Both arguments to the '<tt>shl</tt>' instruction must be the
3827 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3828 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3829
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3830<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3831<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
3832 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
3833 is (statically or dynamically) negative or equal to or larger than the number
3834 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3835 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3836 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3837
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3838<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
3839 <a href="#trapvalues">trap value</a> if it shifts out any non-zero bits. If
Chris Lattner66298c12011-02-09 16:44:44 +0000[diff] [blame]3840 the <tt>nsw</tt> keyword is present, then the shift produces a
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3841 <a href="#trapvalues">trap value</a> if it shifts out any bits that disagree
3842 with the resultant sign bit. As such, NUW/NSW have the same semantics as
3843 they would if the shift were expressed as a mul instruction with the same
3844 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
3845
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3846<h5>Example:</h5>
3847<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3848 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
3849 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
3850 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3851 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]3852 &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]3853</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3854
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3855</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3856
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3857<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3858<h4>
3859 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
3860</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3861
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3862<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3863
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3864<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3865<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3866 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3867 &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]3868</pre>
3869
3870<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3871<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
3872 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3873
3874<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3875<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3876 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3877 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3878
3879<h5>Semantics:</h5>
3880<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3881 significant bits of the result will be filled with zero bits after the shift.
3882 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
3883 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
3884 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
3885 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3886
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3887<p>If the <tt>exact</tt> keyword is present, the result value of the
3888 <tt>lshr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3889 shifted out are non-zero.</p>
3890
3891
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3892<h5>Example:</h5>
3893<pre>
3894 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
3895 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
3896 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
3897 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3898 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]3899 &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]3900</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3901
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3902</div>
3903
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]3904<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3905<h4>
3906 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
3907</h4>
3908
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3909<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3910
3911<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3912<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3913 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3914 &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]3915</pre>
3916
3917<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3918<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
3919 operand shifted to the right a specified number of bits with sign
3920 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3921
3922<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3923<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3924 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3925 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3926
3927<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3928<p>This instruction always performs an arithmetic shift right operation, The
3929 most significant bits of the result will be filled with the sign bit
3930 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
3931 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
3932 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
3933 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3934
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]3935<p>If the <tt>exact</tt> keyword is present, the result value of the
3936 <tt>ashr</tt> is a <a href="#trapvalues">trap value</a> if any of the bits
3937 shifted out are non-zero.</p>
3938
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3939<h5>Example:</h5>
3940<pre>
3941 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
3942 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
3943 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
3944 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]3945 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]3946 &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]3947</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3948
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]3949</div>
3950
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3951<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3952<h4>
3953 <a name="i_and">'<tt>and</tt>' Instruction</a>
3954</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3955
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3956<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3957
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3958<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3959<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3960 &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]3961</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3962
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3963<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3964<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
3965 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3966
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3967<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3968<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3969 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3970 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3971
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3972<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3973<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3974
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3975<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3976 <tbody>
3977 <tr>
3978 <td>In0</td>
3979 <td>In1</td>
3980 <td>Out</td>
3981 </tr>
3982 <tr>
3983 <td>0</td>
3984 <td>0</td>
3985 <td>0</td>
3986 </tr>
3987 <tr>
3988 <td>0</td>
3989 <td>1</td>
3990 <td>0</td>
3991 </tr>
3992 <tr>
3993 <td>1</td>
3994 <td>0</td>
3995 <td>0</td>
3996 </tr>
3997 <tr>
3998 <td>1</td>
3999 <td>1</td>
4000 <td>1</td>
4001 </tr>
4002 </tbody>
4003</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4004
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4005<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4006<pre>
4007 &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]4008 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4009 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4010</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4011</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4012<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4013<h4>
4014 <a name="i_or">'<tt>or</tt>' Instruction</a>
4015</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4016
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4017<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4018
4019<h5>Syntax:</h5>
4020<pre>
4021 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4022</pre>
4023
4024<h5>Overview:</h5>
4025<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4026 two operands.</p>
4027
4028<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4029<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4030 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4031 values. Both arguments must have identical types.</p>
4032
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4033<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4034<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4035
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4036<table border="1" cellspacing="0" cellpadding="4">
4037 <tbody>
4038 <tr>
4039 <td>In0</td>
4040 <td>In1</td>
4041 <td>Out</td>
4042 </tr>
4043 <tr>
4044 <td>0</td>
4045 <td>0</td>
4046 <td>0</td>
4047 </tr>
4048 <tr>
4049 <td>0</td>
4050 <td>1</td>
4051 <td>1</td>
4052 </tr>
4053 <tr>
4054 <td>1</td>
4055 <td>0</td>
4056 <td>1</td>
4057 </tr>
4058 <tr>
4059 <td>1</td>
4060 <td>1</td>
4061 <td>1</td>
4062 </tr>
4063 </tbody>
4064</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4065
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4066<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4067<pre>
4068 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4069 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4070 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4071</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4072
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4073</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4074
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4075<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4076<h4>
4077 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4078</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4079
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4080<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4081
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4082<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4083<pre>
4084 &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]4085</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4086
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4087<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4088<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4089 its two operands. The <tt>xor</tt> is used to implement the "one's
4090 complement" operation, which is the "~" operator in C.</p>
4091
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4092<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4093<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4094 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4095 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4096
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4097<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4098<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4099
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4100<table border="1" cellspacing="0" cellpadding="4">
4101 <tbody>
4102 <tr>
4103 <td>In0</td>
4104 <td>In1</td>
4105 <td>Out</td>
4106 </tr>
4107 <tr>
4108 <td>0</td>
4109 <td>0</td>
4110 <td>0</td>
4111 </tr>
4112 <tr>
4113 <td>0</td>
4114 <td>1</td>
4115 <td>1</td>
4116 </tr>
4117 <tr>
4118 <td>1</td>
4119 <td>0</td>
4120 <td>1</td>
4121 </tr>
4122 <tr>
4123 <td>1</td>
4124 <td>1</td>
4125 <td>0</td>
4126 </tr>
4127 </tbody>
4128</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4129
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4130<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4131<pre>
4132 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4133 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4134 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4135 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4136</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4137
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4138</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4139
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4140</div>
4141
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4142<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4143<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4144 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4145</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4146
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4147<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4148
4149<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4150 target-independent manner. These instructions cover the element-access and
4151 vector-specific operations needed to process vectors effectively. While LLVM
4152 does directly support these vector operations, many sophisticated algorithms
4153 will want to use target-specific intrinsics to take full advantage of a
4154 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4155
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4156<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4157<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4158 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4159</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4160
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4161<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4162
4163<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4164<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4165 &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]4166</pre>
4167
4168<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4169<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4170 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4171
4172
4173<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4174<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4175 of <a href="#t_vector">vector</a> type. The second operand is an index
4176 indicating the position from which to extract the element. The index may be
4177 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4178
4179<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4180<p>The result is a scalar of the same type as the element type of
4181 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4182 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4183 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4184
4185<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4186<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4187 &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]4188</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4189
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4190</div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4191
4192<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4193<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4194 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4195</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4196
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4197<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4198
4199<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4200<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +0000[diff] [blame]4201 &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]4202</pre>
4203
4204<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4205<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4206 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4207
4208<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4209<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4210 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4211 whose type must equal the element type of the first operand. The third
4212 operand is an index indicating the position at which to insert the value.
4213 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4214
4215<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4216<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4217 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4218 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4219 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4220
4221<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4222<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4223 &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]4224</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4225
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4226</div>
4227
4228<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4229<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4230 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4231</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4232
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4233<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4234
4235<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4236<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4237 &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]4238</pre>
4239
4240<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4241<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4242 from two input vectors, returning a vector with the same element type as the
4243 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4244
4245<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4246<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4247 with types that match each other. The third argument is a shuffle mask whose
4248 element type is always 'i32'. The result of the instruction is a vector
4249 whose length is the same as the shuffle mask and whose element type is the
4250 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4251
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4252<p>The shuffle mask operand is required to be a constant vector with either
4253 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4254
4255<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4256<p>The elements of the two input vectors are numbered from left to right across
4257 both of the vectors. The shuffle mask operand specifies, for each element of
4258 the result vector, which element of the two input vectors the result element
4259 gets. The element selector may be undef (meaning "don't care") and the
4260 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4261
4262<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4263<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4264 &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]4265 &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]4266 &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]4267 &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]4268 &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]4269 &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]4270 &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]4271 &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]4272</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4273
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4274</div>
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]4275
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4276</div>
4277
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4278<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4279<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4280 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4281</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4282
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4283<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4284
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4285<p>LLVM supports several instructions for working with
4286 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4287
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4288<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4289<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4290 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4291</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4292
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4293<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4294
4295<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4296<pre>
4297 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4298</pre>
4299
4300<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4301<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4302 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4303
4304<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4305<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4306 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4307 <a href="#t_array">array</a> type. The operands are constant indices to
4308 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4309 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4310 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4311 <ul>
4312 <li>Since the value being indexed is not a pointer, the first index is
4313 omitted and assumed to be zero.</li>
4314 <li>At least one index must be specified.</li>
4315 <li>Not only struct indices but also array indices must be in
4316 bounds.</li>
4317 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4318
4319<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4320<p>The result is the value at the position in the aggregate specified by the
4321 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4322
4323<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4324<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4325 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4326</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4327
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4328</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4329
4330<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4331<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4332 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4333</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4334
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4335<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4336
4337<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4338<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +0000[diff] [blame]4339 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt;{, <idx>}* <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4340</pre>
4341
4342<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4343<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4344 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4345
4346<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4347<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4348 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4349 <a href="#t_array">array</a> type. The second operand is a first-class
4350 value to insert. The following operands are constant indices indicating
4351 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]4352 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4353 value to insert must have the same type as the value identified by the
4354 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4355
4356<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4357<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4358 that of <tt>val</tt> except that the value at the position specified by the
4359 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4360
4361<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4362<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +0000[diff] [blame]4363 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4364 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4365 %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]4366</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4367
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4368</div>
4369
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4370</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4371
4372<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4373<h3>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4374 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4375</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4376
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4377<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4378
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4379<p>A key design point of an SSA-based representation is how it represents
4380 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]4381 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4382 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4383
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4384<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4385<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4386 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4387</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4388
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4389<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4390
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4391<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4392<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +0000[diff] [blame]4393 &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]4394</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4395
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4396<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4397<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4398 currently executing function, to be automatically released when this function
4399 returns to its caller. The object is always allocated in the generic address
4400 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4401
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4402<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4403<p>The '<tt>alloca</tt>' instruction
4404 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4405 runtime stack, returning a pointer of the appropriate type to the program.
4406 If "NumElements" is specified, it is the number of elements allocated,
4407 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4408 specified, the value result of the allocation is guaranteed to be aligned to
4409 at least that boundary. If not specified, or if zero, the target can choose
4410 to align the allocation on any convenient boundary compatible with the
4411 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4412
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4413<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4414
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4415<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +0000[diff] [blame]4416<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4417 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4418 memory is automatically released when the function returns. The
4419 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4420 variables that must have an address available. When the function returns
4421 (either with the <tt><a href="#i_ret">ret</a></tt>
4422 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4423 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4424
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4425<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4426<pre>
Dan Gohman81e21672009-01-04 23:49:44 +0000[diff] [blame]4427 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4428 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4429 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4430 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4431</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4432
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4433</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4434
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4435<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4436<h4>
4437 <a name="i_load">'<tt>load</tt>' Instruction</a>
4438</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4439
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4440<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4441
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4442<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4443<pre>
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4444 &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4445 &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4446 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4447</pre>
4448
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4449<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4450<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4451
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4452<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4453<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4454 from which to load. The pointer must point to
4455 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4456 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4457 number or order of execution of this <tt>load</tt> with other <a
4458 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4459
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4460<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4461 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]4462 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4463 alignment for the target. It is the responsibility of the code emitter to
4464 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4465 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4466 produce less efficient code. An alignment of 1 is always safe.</p>
4467
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4468<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4469 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4470 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4471 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4472 and code generator that this load is not expected to be reused in the cache.
4473 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4474 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4475
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4476<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4477<p>The location of memory pointed to is loaded. If the value being loaded is of
4478 scalar type then the number of bytes read does not exceed the minimum number
4479 of bytes needed to hold all bits of the type. For example, loading an
4480 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4481 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4482 is undefined if the value was not originally written using a store of the
4483 same type.</p>
4484
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4485<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4486<pre>
4487 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4488 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4489 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4490</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4491
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4492</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4493
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4494<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4495<h4>
4496 <a name="i_store">'<tt>store</tt>' Instruction</a>
4497</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4498
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4499<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4500
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4501<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4502<pre>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4503 store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
4504 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4505</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4506
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4507<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4508<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4509
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4510<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4511<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4512 and an address at which to store it. The type of the
4513 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4514 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4515 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4516 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4517 order of execution of this <tt>store</tt> with other <a
4518 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4519
4520<p>The optional constant "align" argument specifies the alignment of the
4521 operation (that is, the alignment of the memory address). A value of 0 or an
4522 omitted "align" argument means that the operation has the preferential
4523 alignment for the target. It is the responsibility of the code emitter to
4524 ensure that the alignment information is correct. Overestimating the
4525 alignment results in an undefined behavior. Underestimating the alignment may
4526 produce less efficient code. An alignment of 1 is always safe.</p>
4527
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4528<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4529 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4530 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4531 instruction tells the optimizer and code generator that this load is
4532 not expected to be reused in the cache. The code generator may
4533 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4534 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4535
4536
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4537<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4538<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4539 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4540 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4541 does not exceed the minimum number of bytes needed to hold all bits of the
4542 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4543 writing a value of a type like <tt>i20</tt> with a size that is not an
4544 integral number of bytes, it is unspecified what happens to the extra bits
4545 that do not belong to the type, but they will typically be overwritten.</p>
4546
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4547<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4548<pre>
4549 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]4550 store i32 3, i32* %ptr <i>; yields {void}</i>
4551 %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]4552</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4553
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]4554</div>
4555
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4556<!-- _______________________________________________________________________ -->
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]4557<div class="doc_subsubsection"> <a name="i_fence">'<tt>fence</tt>'
4558Instruction</a> </div>
4559
4560<div class="doc_text">
4561
4562<h5>Syntax:</h5>
4563<pre>
4564 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
4565</pre>
4566
4567<h5>Overview:</h5>
4568<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
4569between operations.</p>
4570
4571<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
4572href="#ordering">ordering</a> argument which defines what
4573<i>synchronizes-with</i> edges they add. They can only be given
4574<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
4575<code>seq_cst</code> orderings.</p>
4576
4577<h5>Semantics:</h5>
4578<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
4579semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
4580<code>acquire</code> ordering semantics if and only if there exist atomic
4581operations <var>X</var> and <var>Y</var>, both operating on some atomic object
4582<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
4583<var>X</var> modifies <var>M</var> (either directly or through some side effect
4584of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
4585<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
4586<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
4587than an explicit <code>fence</code>, one (but not both) of the atomic operations
4588<var>X</var> or <var>Y</var> might provide a <code>release</code> or
4589<code>acquire</code> (resp.) ordering constraint and still
4590<i>synchronize-with</i> the explicit <code>fence</code> and establish the
4591<i>happens-before</i> edge.</p>
4592
4593<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
4594having both <code>acquire</code> and <code>release</code> semantics specified
4595above, participates in the global program order of other <code>seq_cst</code>
4596operations and/or fences.</p>
4597
4598<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
4599specifies that the fence only synchronizes with other fences in the same
4600thread. (This is useful for interacting with signal handlers.)</p>
4601
4602<p>FIXME: This instruction is a work in progress; until it is finished, use
4603 llvm.memory.barrier.
4604
4605<h5>Example:</h5>
4606<pre>
4607 fence acquire <i>; yields {void}</i>
4608 fence singlethread seq_cst <i>; yields {void}</i>
4609</pre>
4610
4611</div>
4612
4613<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4614<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4615 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4616</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4617
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4618<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4619
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4620<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4621<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4622 &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]4623 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4624</pre>
4625
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4626<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4627<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4628 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
4629 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4630
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4631<h5>Arguments:</h5>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4632<p>The first argument is always a pointer, and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]4633 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4634 elements of the aggregate object are indexed. The interpretation of each
4635 index is dependent on the type being indexed into. The first index always
4636 indexes the pointer value given as the first argument, the second index
4637 indexes a value of the type pointed to (not necessarily the value directly
4638 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4639 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4640 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4641 can never be pointers, since that would require loading the pointer before
4642 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4643
4644<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]4645 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4646 integer <b>constants</b> are allowed. When indexing into an array, pointer
4647 or vector, integers of any width are allowed, and they are not required to be
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]4648 constant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4649
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4650<p>For example, let's consider a C code fragment and how it gets compiled to
4651 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4652
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4653<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4654struct RT {
4655 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4656 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4657 char C;
4658};
4659struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4660 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4661 double Y;
4662 struct RT Z;
4663};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4664
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]4665int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4666 return &amp;s[1].Z.B[5][13];
4667}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4668</pre>
4669
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4670<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4671
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4672<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]4673%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
4674%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4675
Dan Gohman4df605b2009-07-25 02:23:48 +0000[diff] [blame]4676define i32* @foo(%ST* %s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]4677entry:
4678 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
4679 ret i32* %reg
4680}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4681</pre>
4682
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4683<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4684<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4685 type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
4686 }</tt>' type, a structure. The second index indexes into the third element
4687 of the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
4688 i8 }</tt>' type, another structure. The third index indexes into the second
4689 element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
4690 array. The two dimensions of the array are subscripted into, yielding an
4691 '<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a
4692 pointer to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4693
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4694<p>Note that it is perfectly legal to index partially through a structure,
4695 returning a pointer to an inner element. Because of this, the LLVM code for
4696 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4697
4698<pre>
Dan Gohman4df605b2009-07-25 02:23:48 +0000[diff] [blame]4699 define i32* @foo(%ST* %s) {
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4700 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4701 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
4702 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4703 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
4704 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
4705 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4706 }
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]4707</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]4708
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]4709<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]4710 <tt>getelementptr</tt> is a <a href="#trapvalues">trap value</a> if the
4711 base pointer is not an <i>in bounds</i> address of an allocated object,
4712 or if any of the addresses that would be formed by successive addition of
4713 the offsets implied by the indices to the base address with infinitely
4714 precise arithmetic are not an <i>in bounds</i> address of that allocated
4715 object. The <i>in bounds</i> addresses for an allocated object are all
4716 the addresses that point into the object, plus the address one byte past
4717 the end.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]4718
4719<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
4720 the base address with silently-wrapping two's complement arithmetic, and
4721 the result value of the <tt>getelementptr</tt> may be outside the object
4722 pointed to by the base pointer. The result value may not necessarily be
4723 used to access memory though, even if it happens to point into allocated
4724 storage. See the <a href="#pointeraliasing">Pointer Aliasing Rules</a>
4725 section for more information.</p>
4726
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4727<p>The getelementptr instruction is often confusing. For some more insight into
4728 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4729
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]4730<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4731<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4732 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]4733 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
4734 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]4735 %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]4736 <i>; yields i8*:eptr</i>
4737 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +0000[diff] [blame]4738 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +0000[diff] [blame]4739 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4740</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4741
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]4742</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]4743
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4744</div>
4745
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4746<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4747<h3>
4748 <a name="convertops">Conversion Operations</a>
4749</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4750
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4751<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4752
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]4753<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4754 which all take a single operand and a type. They perform various bit
4755 conversions on the operand.</p>
4756
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]4757<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4758<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4759 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4760</h4>
4761
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4762<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4763
4764<h5>Syntax:</h5>
4765<pre>
4766 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4767</pre>
4768
4769<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4770<p>The '<tt>trunc</tt>' instruction truncates its operand to the
4771 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4772
4773<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4774<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
4775 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4776 of the same number of integers.
4777 The bit size of the <tt>value</tt> must be larger than
4778 the bit size of the destination type, <tt>ty2</tt>.
4779 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4780
4781<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4782<p>The '<tt>trunc</tt>' instruction truncates the high order bits
4783 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
4784 source size must be larger than the destination size, <tt>trunc</tt> cannot
4785 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4786
4787<h5>Example:</h5>
4788<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4789 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
4790 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
4791 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
4792 %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]4793</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4794
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4795</div>
4796
4797<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4798<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4799 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4800</h4>
4801
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4802<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4803
4804<h5>Syntax:</h5>
4805<pre>
4806 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4807</pre>
4808
4809<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4810<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4811 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4812
4813
4814<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]4815<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
4816 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4817 of the same number of integers.
4818 The bit size of the <tt>value</tt> must be smaller than
4819 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4820 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4821
4822<h5>Semantics:</h5>
4823<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]4824 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4825
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]4826<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]4827
4828<h5>Example:</h5>
4829<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4830 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]4831 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]4832 %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]4833</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4834
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4835</div>
4836
4837<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4838<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4839 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4840</h4>
4841
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4842<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4843
4844<h5>Syntax:</h5>
4845<pre>
4846 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4847</pre>
4848
4849<h5>Overview:</h5>
4850<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
4851
4852<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4853<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
4854 Both types must be of <a href="#t_integer">integer</a> types, or vectors
4855 of the same number of integers.
4856 The bit size of the <tt>value</tt> must be smaller than
4857 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4858 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4859
4860<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4861<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
4862 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
4863 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4864
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]4865<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]4866
4867<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4868<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4869 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]4870 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]4871 %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]4872</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4873
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4874</div>
4875
4876<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4877<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4878 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4879</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4880
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4881<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4882
4883<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4884<pre>
4885 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4886</pre>
4887
4888<h5>Overview:</h5>
4889<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4890 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4891
4892<h5>Arguments:</h5>
4893<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4894 point</a> value to cast and a <a href="#t_floating">floating point</a> type
4895 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]4896 <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]4897 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4898
4899<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4900<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4901 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4902 <a href="#t_floating">floating point</a> type. If the value cannot fit
4903 within the destination type, <tt>ty2</tt>, then the results are
4904 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4905
4906<h5>Example:</h5>
4907<pre>
4908 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
4909 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
4910</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4911
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]4912</div>
4913
4914<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4915<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4916 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4917</h4>
4918
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4919<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4920
4921<h5>Syntax:</h5>
4922<pre>
4923 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
4924</pre>
4925
4926<h5>Overview:</h5>
4927<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]4928 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4929
4930<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4931<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4932 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
4933 a <a href="#t_floating">floating point</a> type to cast it to. The source
4934 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4935
4936<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4937<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4938 <a href="#t_floating">floating point</a> type to a larger
4939 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
4940 used to make a <i>no-op cast</i> because it always changes bits. Use
4941 <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]4942
4943<h5>Example:</h5>
4944<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +0000[diff] [blame]4945 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
4946 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4947</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4948
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4949</div>
4950
4951<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4952<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]4953 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4954</h4>
4955
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4956<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4957
4958<h5>Syntax:</h5>
4959<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]4960 &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]4961</pre>
4962
4963<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]4964<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4965 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4966
4967<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4968<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
4969 scalar or vector <a href="#t_floating">floating point</a> value, and a type
4970 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
4971 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
4972 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4973
4974<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4975<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4976 <a href="#t_floating">floating point</a> operand into the nearest (rounding
4977 towards zero) unsigned integer value. If the value cannot fit
4978 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4979
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4980<h5>Example:</h5>
4981<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]4982 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]4983 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]4984 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4985</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4986
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4987</div>
4988
4989<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4990<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4991 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4992</h4>
4993
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4994<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]4995
4996<h5>Syntax:</h5>
4997<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]4998 &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]4999</pre>
5000
5001<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5002<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5003 <a href="#t_floating">floating point</a> <tt>value</tt> to
5004 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5005
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5006<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5007<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5008 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5009 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5010 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5011 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5012
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5013<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5014<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5015 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5016 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5017 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5018
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5019<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5020<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5021 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5022 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5023 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5024</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5025
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5026</div>
5027
5028<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5029<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5030 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5031</h4>
5032
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5033<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5034
5035<h5>Syntax:</h5>
5036<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5037 &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]5038</pre>
5039
5040<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5041<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5042 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5043
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5044<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5045<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]5046 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5047 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5048 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5049 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5050
5051<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5052<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5053 integer quantity and converts it to the corresponding floating point
5054 value. If the value cannot fit in the floating point value, the results are
5055 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5056
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5057<h5>Example:</h5>
5058<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5059 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5060 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5061</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5062
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5063</div>
5064
5065<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5066<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5067 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5068</h4>
5069
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5070<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5071
5072<h5>Syntax:</h5>
5073<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5074 &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]5075</pre>
5076
5077<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5078<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5079 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5080
5081<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5082<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]5083 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5084 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5085 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5086 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5087
5088<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5089<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5090 quantity and converts it to the corresponding floating point value. If the
5091 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5092
5093<h5>Example:</h5>
5094<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5095 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5096 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5097</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5098
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5099</div>
5100
5101<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5102<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5103 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5104</h4>
5105
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5106<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5107
5108<h5>Syntax:</h5>
5109<pre>
5110 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5111</pre>
5112
5113<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5114<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
5115 the integer type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5116
5117<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5118<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
5119 must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
5120 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5121
5122<h5>Semantics:</h5>
5123<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5124 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5125 truncating or zero extending that value to the size of the integer type. If
5126 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5127 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5128 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5129 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5130
5131<h5>Example:</h5>
5132<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5133 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
5134 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5135</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5136
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5137</div>
5138
5139<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5140<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5141 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5142</h4>
5143
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5144<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5145
5146<h5>Syntax:</h5>
5147<pre>
5148 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5149</pre>
5150
5151<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5152<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5153 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5154
5155<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]5156<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5157 value to cast, and a type to cast it to, which must be a
5158 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5159
5160<h5>Semantics:</h5>
5161<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5162 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5163 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5164 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5165 than the size of a pointer then a zero extension is done. If they are the
5166 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5167
5168<h5>Example:</h5>
5169<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5170 %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]5171 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5172 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5173</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5174
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5175</div>
5176
5177<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5178<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5179 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5180</h4>
5181
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5182<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5183
5184<h5>Syntax:</h5>
5185<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5186 &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]5187</pre>
5188
5189<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5190<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5191 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5192
5193<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5194<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5195 non-aggregate first class value, and a type to cast it to, which must also be
5196 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5197 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5198 identical. If the source type is a pointer, the destination type must also be
5199 a pointer. This instruction supports bitwise conversion of vectors to
5200 integers and to vectors of other types (as long as they have the same
5201 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5202
5203<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5204<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5205 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5206 this conversion. The conversion is done as if the <tt>value</tt> had been
5207 stored to memory and read back as type <tt>ty2</tt>. Pointer types may only
5208 be converted to other pointer types with this instruction. To convert
5209 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5210 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5211
5212<h5>Example:</h5>
5213<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5214 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5215 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5216 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5217</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5218
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5219</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5220
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5221</div>
5222
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5223<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5224<h3>
5225 <a name="otherops">Other Operations</a>
5226</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5227
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5228<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5229
5230<p>The instructions in this category are the "miscellaneous" instructions, which
5231 defy better classification.</p>
5232
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5233<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5234<h4>
5235 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5236</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5237
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5238<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5239
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5240<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5241<pre>
5242 &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]5243</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5244
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5245<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5246<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
5247 boolean values based on comparison of its two integer, integer vector, or
5248 pointer operands.</p>
5249
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5250<h5>Arguments:</h5>
5251<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5252 the condition code indicating the kind of comparison to perform. It is not a
5253 value, just a keyword. The possible condition code are:</p>
5254
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5255<ol>
5256 <li><tt>eq</tt>: equal</li>
5257 <li><tt>ne</tt>: not equal </li>
5258 <li><tt>ugt</tt>: unsigned greater than</li>
5259 <li><tt>uge</tt>: unsigned greater or equal</li>
5260 <li><tt>ult</tt>: unsigned less than</li>
5261 <li><tt>ule</tt>: unsigned less or equal</li>
5262 <li><tt>sgt</tt>: signed greater than</li>
5263 <li><tt>sge</tt>: signed greater or equal</li>
5264 <li><tt>slt</tt>: signed less than</li>
5265 <li><tt>sle</tt>: signed less or equal</li>
5266</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5267
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]5268<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5269 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5270 typed. They must also be identical types.</p>
5271
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5272<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5273<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5274 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5275 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]5276 result, as follows:</p>
5277
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5278<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5279 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5280 <tt>false</tt> otherwise. No sign interpretation is necessary or
5281 performed.</li>
5282
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5283 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5284 <tt>false</tt> otherwise. No sign interpretation is necessary or
5285 performed.</li>
5286
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5287 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5288 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5289
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5290 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5291 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5292 to <tt>op2</tt>.</li>
5293
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5294 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5295 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5296
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5297 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5298 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5299
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5300 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5301 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5302
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5303 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5304 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5305 to <tt>op2</tt>.</li>
5306
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5307 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5308 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5309
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5310 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5311 <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]5312</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5313
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5314<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5315 values are compared as if they were integers.</p>
5316
5317<p>If the operands are integer vectors, then they are compared element by
5318 element. The result is an <tt>i1</tt> vector with the same number of elements
5319 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5320
5321<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5322<pre>
5323 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5324 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5325 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5326 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5327 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5328 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5329</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5330
5331<p>Note that the code generator does not yet support vector types with
5332 the <tt>icmp</tt> instruction.</p>
5333
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5334</div>
5335
5336<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5337<h4>
5338 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5339</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5340
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5341<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5342
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5343<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5344<pre>
5345 &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]5346</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5347
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5348<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5349<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5350 values based on comparison of its operands.</p>
5351
5352<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]5353(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5354
5355<p>If the operands are floating point vectors, then the result type is a vector
5356 of boolean with the same number of elements as the operands being
5357 compared.</p>
5358
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5359<h5>Arguments:</h5>
5360<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5361 the condition code indicating the kind of comparison to perform. It is not a
5362 value, just a keyword. The possible condition code are:</p>
5363
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5364<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5365 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5366 <li><tt>oeq</tt>: ordered and equal</li>
5367 <li><tt>ogt</tt>: ordered and greater than </li>
5368 <li><tt>oge</tt>: ordered and greater than or equal</li>
5369 <li><tt>olt</tt>: ordered and less than </li>
5370 <li><tt>ole</tt>: ordered and less than or equal</li>
5371 <li><tt>one</tt>: ordered and not equal</li>
5372 <li><tt>ord</tt>: ordered (no nans)</li>
5373 <li><tt>ueq</tt>: unordered or equal</li>
5374 <li><tt>ugt</tt>: unordered or greater than </li>
5375 <li><tt>uge</tt>: unordered or greater than or equal</li>
5376 <li><tt>ult</tt>: unordered or less than </li>
5377 <li><tt>ule</tt>: unordered or less than or equal</li>
5378 <li><tt>une</tt>: unordered or not equal</li>
5379 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5380 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5381</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5382
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5383<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5384 <i>unordered</i> means that either operand may be a QNAN.</p>
5385
5386<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5387 a <a href="#t_floating">floating point</a> type or
5388 a <a href="#t_vector">vector</a> of floating point type. They must have
5389 identical types.</p>
5390
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5391<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]5392<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5393 according to the condition code given as <tt>cond</tt>. If the operands are
5394 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5395 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5396 follows:</p>
5397
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5398<ol>
5399 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5400
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5401 <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]5402 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5403
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5404 <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]5405 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5406
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5407 <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]5408 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5409
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5410 <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]5411 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5412
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5413 <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]5414 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5415
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5416 <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]5417 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5418
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5419 <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]5420
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5421 <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]5422 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5423
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5424 <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]5425 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5426
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5427 <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]5428 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5429
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5430 <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]5431 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5432
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5433 <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]5434 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5435
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5436 <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]5437 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5438
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5439 <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]5440
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5441 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5442</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5443
5444<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5445<pre>
5446 &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]5447 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5448 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5449 &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]5450</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5451
5452<p>Note that the code generator does not yet support vector types with
5453 the <tt>fcmp</tt> instruction.</p>
5454
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5455</div>
5456
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5457<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5458<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5459 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5460</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5461
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5462<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5463
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5464<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5465<pre>
5466 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5467</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5468
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5469<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5470<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5471 SSA graph representing the function.</p>
5472
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5473<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5474<p>The type of the incoming values is specified with the first type field. After
5475 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5476 one pair for each predecessor basic block of the current block. Only values
5477 of <a href="#t_firstclass">first class</a> type may be used as the value
5478 arguments to the PHI node. Only labels may be used as the label
5479 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5480
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5481<p>There must be no non-phi instructions between the start of a basic block and
5482 the PHI instructions: i.e. PHI instructions must be first in a basic
5483 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5484
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5485<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5486 occur on the edge from the corresponding predecessor block to the current
5487 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5488 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +0000[diff] [blame]5489
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5490<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5491<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5492 specified by the pair corresponding to the predecessor basic block that
5493 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5494
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5495<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5496<pre>
5497Loop: ; Infinite loop that counts from 0 on up...
5498 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5499 %nextindvar = add i32 %indvar, 1
5500 br label %Loop
5501</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5502
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5503</div>
5504
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5505<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5506<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5507 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5508</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5509
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5510<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5511
5512<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5513<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]5514 &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>
5515
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5516 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5517</pre>
5518
5519<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5520<p>The '<tt>select</tt>' instruction is used to choose one value based on a
5521 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5522
5523
5524<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5525<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
5526 values indicating the condition, and two values of the
5527 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
5528 vectors and the condition is a scalar, then entire vectors are selected, not
5529 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5530
5531<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5532<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
5533 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5534
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5535<p>If the condition is a vector of i1, then the value arguments must be vectors
5536 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5537
5538<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5539<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5540 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5541</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5542
5543<p>Note that the code generator does not yet support conditions
5544 with vector type.</p>
5545
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5546</div>
5547
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]5548<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5549<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5550 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5551</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5552
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5553<div>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5554
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5555<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5556<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]5557 &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]5558</pre>
5559
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5560<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5561<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5562
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5563<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5564<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5565
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5566<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5567 <li>The optional "tail" marker indicates that the callee function does not
5568 access any allocas or varargs in the caller. Note that calls may be
5569 marked "tail" even if they do not occur before
5570 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
5571 present, the function call is eligible for tail call optimization,
5572 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +0000[diff] [blame]5573 optimized into a jump</a>. The code generator may optimize calls marked
5574 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
5575 sibling call optimization</a> when the caller and callee have
5576 matching signatures, or 2) forced tail call optimization when the
5577 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5578 <ul>
5579 <li>Caller and callee both have the calling
5580 convention <tt>fastcc</tt>.</li>
5581 <li>The call is in tail position (ret immediately follows call and ret
5582 uses value of call or is void).</li>
5583 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +0000[diff] [blame]5584 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5585 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
5586 constraints are met.</a></li>
5587 </ul>
5588 </li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5589
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5590 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
5591 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]5592 defaults to using C calling conventions. The calling convention of the
5593 call must match the calling convention of the target function, or else the
5594 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5595
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5596 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
5597 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
5598 '<tt>inreg</tt>' attributes are valid here.</li>
5599
5600 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
5601 type of the return value. Functions that return no value are marked
5602 <tt><a href="#t_void">void</a></tt>.</li>
5603
5604 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
5605 being invoked. The argument types must match the types implied by this
5606 signature. This type can be omitted if the function is not varargs and if
5607 the function type does not return a pointer to a function.</li>
5608
5609 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
5610 be invoked. In most cases, this is a direct function invocation, but
5611 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
5612 to function value.</li>
5613
5614 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]5615 signature argument types and parameter attributes. All arguments must be
5616 of <a href="#t_firstclass">first class</a> type. If the function
5617 signature indicates the function accepts a variable number of arguments,
5618 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5619
5620 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
5621 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
5622 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5623</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5624
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5625<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5626<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
5627 a specified function, with its incoming arguments bound to the specified
5628 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
5629 function, control flow continues with the instruction after the function
5630 call, and the return value of the function is bound to the result
5631 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5632
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5633<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5634<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]5635 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]5636 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +0000[diff] [blame]5637 %X = tail call i32 @foo() <i>; yields i32</i>
5638 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
5639 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]5640
5641 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]5642 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]5643 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
5644 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +0000[diff] [blame]5645 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +0000[diff] [blame]5646 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]5647</pre>
5648
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]5649<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +0000[diff] [blame]5650standard C99 library as being the C99 library functions, and may perform
5651optimizations or generate code for them under that assumption. This is
5652something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5653freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]5654
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5655</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5656
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5657<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5658<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]5659 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5660</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5661
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5662<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5663
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5664<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5665<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5666 &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]5667</pre>
5668
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5669<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5670<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5671 the "variable argument" area of a function call. It is used to implement the
5672 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5673
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5674<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5675<p>This instruction takes a <tt>va_list*</tt> value and the type of the
5676 argument. It returns a value of the specified argument type and increments
5677 the <tt>va_list</tt> to point to the next argument. The actual type
5678 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5679
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5680<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5681<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
5682 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
5683 to the next argument. For more information, see the variable argument
5684 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5685
5686<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]5687 take a variable number of arguments, for example, the <tt>vfprintf</tt>
5688 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5689
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5690<p><tt>va_arg</tt> is an LLVM instruction instead of
5691 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
5692 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5693
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]5694<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]5695<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
5696
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5697<p>Note that the code generator does not yet fully support va_arg on many
5698 targets. Also, it does not currently support va_arg with aggregate types on
5699 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]5700
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5701</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5702
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5703</div>
5704
5705</div>
5706
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5707<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5708<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]5709<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5710
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5711<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5712
5713<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5714 well known names and semantics and are required to follow certain
5715 restrictions. Overall, these intrinsics represent an extension mechanism for
5716 the LLVM language that does not require changing all of the transformations
5717 in LLVM when adding to the language (or the bitcode reader/writer, the
5718 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5719
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]5720<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5721 prefix is reserved in LLVM for intrinsic names; thus, function names may not
5722 begin with this prefix. Intrinsic functions must always be external
5723 functions: you cannot define the body of intrinsic functions. Intrinsic
5724 functions may only be used in call or invoke instructions: it is illegal to
5725 take the address of an intrinsic function. Additionally, because intrinsic
5726 functions are part of the LLVM language, it is required if any are added that
5727 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5728
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5729<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
5730 family of functions that perform the same operation but on different data
5731 types. Because LLVM can represent over 8 million different integer types,
5732 overloading is used commonly to allow an intrinsic function to operate on any
5733 integer type. One or more of the argument types or the result type can be
5734 overloaded to accept any integer type. Argument types may also be defined as
5735 exactly matching a previous argument's type or the result type. This allows
5736 an intrinsic function which accepts multiple arguments, but needs all of them
5737 to be of the same type, to only be overloaded with respect to a single
5738 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5739
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5740<p>Overloaded intrinsics will have the names of its overloaded argument types
5741 encoded into its function name, each preceded by a period. Only those types
5742 which are overloaded result in a name suffix. Arguments whose type is matched
5743 against another type do not. For example, the <tt>llvm.ctpop</tt> function
5744 can take an integer of any width and returns an integer of exactly the same
5745 integer width. This leads to a family of functions such as
5746 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
5747 %val)</tt>. Only one type, the return type, is overloaded, and only one type
5748 suffix is required. Because the argument's type is matched against the return
5749 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]5750
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5751<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5752 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5753
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5754<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5755<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5756 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5757</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5758
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5759<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5760
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5761<p>Variable argument support is defined in LLVM with
5762 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
5763 intrinsic functions. These functions are related to the similarly named
5764 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5765
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5766<p>All of these functions operate on arguments that use a target-specific value
5767 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
5768 not define what this type is, so all transformations should be prepared to
5769 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5770
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]5771<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]5772 instruction and the variable argument handling intrinsic functions are
5773 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5774
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5775<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5776define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5777 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5778 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5779 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5780 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5781
5782 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5783 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5784
5785 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5786 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5787 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5788 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5789 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5790
5791 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5792 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5793 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5794}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]5795
5796declare void @llvm.va_start(i8*)
5797declare void @llvm.va_copy(i8*, i8*)
5798declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5799</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5800
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5801<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5802<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5803 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5804</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5805
5806
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5807<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5808
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5809<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5810<pre>
5811 declare void %llvm.va_start(i8* &lt;arglist&gt;)
5812</pre>
5813
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5814<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5815<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
5816 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]5817
5818<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5819<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]5820
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5821<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5822<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]5823 macro available in C. In a target-dependent way, it initializes
5824 the <tt>va_list</tt> element to which the argument points, so that the next
5825 call to <tt>va_arg</tt> will produce the first variable argument passed to
5826 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
5827 need to know the last argument of the function as the compiler can figure
5828 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5829
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5830</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5831
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5832<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5833<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5834 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5835</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5836
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5837<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5838
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5839<h5>Syntax:</h5>
5840<pre>
5841 declare void @llvm.va_end(i8* &lt;arglist&gt;)
5842</pre>
5843
5844<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5845<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5846 which has been initialized previously
5847 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
5848 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5849
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5850<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5851<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]5852
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5853<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5854<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]5855 macro available in C. In a target-dependent way, it destroys
5856 the <tt>va_list</tt> element to which the argument points. Calls
5857 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
5858 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
5859 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]5860
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5861</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5862
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5863<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5864<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5865 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5866</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5867
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5868<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5869
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5870<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5871<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]5872 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5873</pre>
5874
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5875<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5876<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5877 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5878
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5879<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]5880<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]5881 The second argument is a pointer to a <tt>va_list</tt> element to copy
5882 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5883
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]5884<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5885<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]5886 macro available in C. In a target-dependent way, it copies the
5887 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
5888 element. This intrinsic is necessary because
5889 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
5890 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5891
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5892</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]5893
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5894</div>
5895
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]5896<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5897<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5898 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5899</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5900
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5901<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5902
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5903<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]5904Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5905intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
5906roots on the stack</a>, as well as garbage collector implementations that
5907require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
5908barriers. Front-ends for type-safe garbage collected languages should generate
5909these intrinsics to make use of the LLVM garbage collectors. For more details,
5910see <a href="GarbageCollection.html">Accurate Garbage Collection with
5911LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]5912
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5913<p>The garbage collection intrinsics only operate on objects in the generic
5914 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]5915
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5916<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5917<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5918 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5919</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5920
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5921<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5922
5923<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5924<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5925 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5926</pre>
5927
5928<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]5929<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]5930 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5931
5932<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5933<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5934 root pointer. The second pointer (which must be either a constant or a
5935 global value address) contains the meta-data to be associated with the
5936 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5937
5938<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]5939<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]5940 location. At compile-time, the code generator generates information to allow
5941 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
5942 intrinsic may only be used in a function which <a href="#gc">specifies a GC
5943 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5944
5945</div>
5946
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5947<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5948<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5949 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5950</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5951
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5952<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5953
5954<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5955<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5956 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5957</pre>
5958
5959<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5960<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5961 locations, allowing garbage collector implementations that require read
5962 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5963
5964<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]5965<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]5966 allocated from the garbage collector. The first object is a pointer to the
5967 start of the referenced object, if needed by the language runtime (otherwise
5968 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5969
5970<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5971<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5972 instruction, but may be replaced with substantially more complex code by the
5973 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
5974 may only be used in a function which <a href="#gc">specifies a GC
5975 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5976
5977</div>
5978
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5979<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5980<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]5981 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5982</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5983
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5984<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5985
5986<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5987<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]5988 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5989</pre>
5990
5991<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5992<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5993 locations, allowing garbage collector implementations that require write
5994 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]5995
5996<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]5997<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]5998 object to store it to, and the third is the address of the field of Obj to
5999 store to. If the runtime does not require a pointer to the object, Obj may
6000 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6001
6002<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6003<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6004 instruction, but may be replaced with substantially more complex code by the
6005 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6006 may only be used in a function which <a href="#gc">specifies a GC
6007 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6008
6009</div>
6010
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6011</div>
6012
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6013<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6014<h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6015 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6016</h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6017
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6018<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6019
6020<p>These intrinsics are provided by LLVM to expose special features that may
6021 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6022
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6023<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6024<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6025 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6026</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6027
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6028<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6029
6030<h5>Syntax:</h5>
6031<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6032 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6033</pre>
6034
6035<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6036<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6037 target-specific value indicating the return address of the current function
6038 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6039
6040<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6041<p>The argument to this intrinsic indicates which function to return the address
6042 for. Zero indicates the calling function, one indicates its caller, etc.
6043 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6044
6045<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6046<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6047 indicating the return address of the specified call frame, or zero if it
6048 cannot be identified. The value returned by this intrinsic is likely to be
6049 incorrect or 0 for arguments other than zero, so it should only be used for
6050 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6051
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6052<p>Note that calling this intrinsic does not prevent function inlining or other
6053 aggressive transformations, so the value returned may not be that of the
6054 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6055
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6056</div>
6057
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6058<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6059<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6060 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6061</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6062
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6063<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6064
6065<h5>Syntax:</h5>
6066<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6067 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6068</pre>
6069
6070<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6071<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6072 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6073
6074<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6075<p>The argument to this intrinsic indicates which function to return the frame
6076 pointer for. Zero indicates the calling function, one indicates its caller,
6077 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6078
6079<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6080<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6081 indicating the frame address of the specified call frame, or zero if it
6082 cannot be identified. The value returned by this intrinsic is likely to be
6083 incorrect or 0 for arguments other than zero, so it should only be used for
6084 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6085
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6086<p>Note that calling this intrinsic does not prevent function inlining or other
6087 aggressive transformations, so the value returned may not be that of the
6088 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6089
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6090</div>
6091
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6092<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6093<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6094 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6095</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6096
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6097<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6098
6099<h5>Syntax:</h5>
6100<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6101 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6102</pre>
6103
6104<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6105<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6106 of the function stack, for use
6107 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6108 useful for implementing language features like scoped automatic variable
6109 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6110
6111<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6112<p>This intrinsic returns a opaque pointer value that can be passed
6113 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6114 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6115 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6116 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6117 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6118 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6119
6120</div>
6121
6122<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6123<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6124 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6125</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6126
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6127<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6128
6129<h5>Syntax:</h5>
6130<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6131 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6132</pre>
6133
6134<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6135<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6136 the function stack to the state it was in when the
6137 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6138 executed. This is useful for implementing language features like scoped
6139 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6140
6141<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6142<p>See the description
6143 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6144
6145</div>
6146
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6147<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6148<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6149 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6150</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6151
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6152<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6153
6154<h5>Syntax:</h5>
6155<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6156 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]6157</pre>
6158
6159<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6160<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6161 insert a prefetch instruction if supported; otherwise, it is a noop.
6162 Prefetches have no effect on the behavior of the program but can change its
6163 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6164
6165<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6166<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6167 specifier determining if the fetch should be for a read (0) or write (1),
6168 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6169 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6170 specifies whether the prefetch is performed on the data (1) or instruction (0)
6171 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6172 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6173
6174<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6175<p>This intrinsic does not modify the behavior of the program. In particular,
6176 prefetches cannot trap and do not produce a value. On targets that support
6177 this intrinsic, the prefetch can provide hints to the processor cache for
6178 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6179
6180</div>
6181
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6182<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6183<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6184 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6185</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6186
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6187<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6188
6189<h5>Syntax:</h5>
6190<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6191 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6192</pre>
6193
6194<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6195<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
6196 Counter (PC) in a region of code to simulators and other tools. The method
6197 is target specific, but it is expected that the marker will use exported
6198 symbols to transmit the PC of the marker. The marker makes no guarantees
6199 that it will remain with any specific instruction after optimizations. It is
6200 possible that the presence of a marker will inhibit optimizations. The
6201 intended use is to be inserted after optimizations to allow correlations of
6202 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6203
6204<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6205<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6206
6207<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6208<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6209 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6210
6211</div>
6212
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6213<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6214<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6215 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6216</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6217
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6218<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6219
6220<h5>Syntax:</h5>
6221<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]6222 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6223</pre>
6224
6225<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6226<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6227 counter register (or similar low latency, high accuracy clocks) on those
6228 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6229 should map to RPCC. As the backing counters overflow quickly (on the order
6230 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6231
6232<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6233<p>When directly supported, reading the cycle counter should not modify any
6234 memory. Implementations are allowed to either return a application specific
6235 value or a system wide value. On backends without support, this is lowered
6236 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6237
6238</div>
6239
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6240</div>
6241
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6242<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6243<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6244 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6245</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6246
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6247<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6248
6249<p>LLVM provides intrinsics for a few important standard C library functions.
6250 These intrinsics allow source-language front-ends to pass information about
6251 the alignment of the pointer arguments to the code generator, providing
6252 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6253
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6254<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6255<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6256 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6257</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6258
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6259<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6260
6261<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6262<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]6263 integer bit width and for different address spaces. Not all targets support
6264 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6265
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6266<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6267 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]6268 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6269 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]6270 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6271</pre>
6272
6273<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6274<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6275 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6276
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6277<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6278 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6279 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6280
6281<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6282
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6283<p>The first argument is a pointer to the destination, the second is a pointer
6284 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6285 number of bytes to copy, the fourth argument is the alignment of the
6286 source and destination locations, and the fifth is a boolean indicating a
6287 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6288
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6289<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]6290 then the caller guarantees that both the source and destination pointers are
6291 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6292
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6293<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6294 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6295 The detailed access behavior is not very cleanly specified and it is unwise
6296 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6297
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6298<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6299
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6300<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6301 source location to the destination location, which are not allowed to
6302 overlap. It copies "len" bytes of memory over. If the argument is known to
6303 be aligned to some boundary, this can be specified as the fourth argument,
6304 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6305
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6306</div>
6307
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6308<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6309<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6310 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6311</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6312
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6313<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6314
6315<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6316<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]6317 width and for different address space. Not all targets support all bit
6318 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6319
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6320<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6321 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]6322 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6323 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]6324 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6325</pre>
6326
6327<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6328<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6329 source location to the destination location. It is similar to the
6330 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6331 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6332
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6333<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6334 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6335 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6336
6337<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6338
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6339<p>The first argument is a pointer to the destination, the second is a pointer
6340 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6341 number of bytes to copy, the fourth argument is the alignment of the
6342 source and destination locations, and the fifth is a boolean indicating a
6343 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6344
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6345<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]6346 then the caller guarantees that the source and destination pointers are
6347 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6348
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6349<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6350 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6351 The detailed access behavior is not very cleanly specified and it is unwise
6352 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6353
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6354<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6355
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6356<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6357 source location to the destination location, which may overlap. It copies
6358 "len" bytes of memory over. If the argument is known to be aligned to some
6359 boundary, this can be specified as the fourth argument, otherwise it should
6360 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6361
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6362</div>
6363
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6364<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6365<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6366 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6367</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6368
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6369<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6370
6371<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6372<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]6373 width and for different address spaces. However, not all targets support all
6374 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6375
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6376<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6377 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6378 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6379 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6380 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6381</pre>
6382
6383<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6384<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6385 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6386
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6387<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6388 intrinsic does not return a value and takes extra alignment/volatile
6389 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6390
6391<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6392<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]6393 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6394 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6395 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6396
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6397<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]6398 then the caller guarantees that the destination pointer is aligned to that
6399 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6400
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6401<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6402 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6403 The detailed access behavior is not very cleanly specified and it is unwise
6404 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6405
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6406<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6407<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6408 at the destination location. If the argument is known to be aligned to some
6409 boundary, this can be specified as the fourth argument, otherwise it should
6410 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6411
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6412</div>
6413
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]6414<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6415<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6416 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6417</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6418
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6419<div>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6420
6421<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6422<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6423 floating point or vector of floating point type. Not all targets support all
6424 types however.</p>
6425
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6426<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]6427 declare float @llvm.sqrt.f32(float %Val)
6428 declare double @llvm.sqrt.f64(double %Val)
6429 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6430 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6431 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6432</pre>
6433
6434<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6435<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6436 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6437 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
6438 behavior for negative numbers other than -0.0 (which allows for better
6439 optimization, because there is no need to worry about errno being
6440 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]6441
6442<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6443<p>The argument and return value are floating point numbers of the same
6444 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6445
6446<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6447<p>This function returns the sqrt of the specified operand if it is a
6448 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6449
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6450</div>
6451
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6452<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6453<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6454 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6455</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6456
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6457<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6458
6459<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6460<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
6461 floating point or vector of floating point type. Not all targets support all
6462 types however.</p>
6463
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6464<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]6465 declare float @llvm.powi.f32(float %Val, i32 %power)
6466 declare double @llvm.powi.f64(double %Val, i32 %power)
6467 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
6468 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
6469 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6470</pre>
6471
6472<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6473<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
6474 specified (positive or negative) power. The order of evaluation of
6475 multiplications is not defined. When a vector of floating point type is
6476 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6477
6478<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6479<p>The second argument is an integer power, and the first is a value to raise to
6480 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6481
6482<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6483<p>This function returns the first value raised to the second power with an
6484 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6485
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]6486</div>
6487
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6488<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6489<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6490 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6491</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6492
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6493<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6494
6495<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6496<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
6497 floating point or vector of floating point type. Not all targets support all
6498 types however.</p>
6499
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6500<pre>
6501 declare float @llvm.sin.f32(float %Val)
6502 declare double @llvm.sin.f64(double %Val)
6503 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
6504 declare fp128 @llvm.sin.f128(fp128 %Val)
6505 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
6506</pre>
6507
6508<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6509<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6510
6511<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6512<p>The argument and return value are floating point numbers of the same
6513 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6514
6515<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6516<p>This function returns the sine of the specified operand, returning the same
6517 values as the libm <tt>sin</tt> functions would, and handles error conditions
6518 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6519
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6520</div>
6521
6522<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6523<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6524 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6525</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6526
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6527<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6528
6529<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6530<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
6531 floating point or vector of floating point type. Not all targets support all
6532 types however.</p>
6533
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6534<pre>
6535 declare float @llvm.cos.f32(float %Val)
6536 declare double @llvm.cos.f64(double %Val)
6537 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
6538 declare fp128 @llvm.cos.f128(fp128 %Val)
6539 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
6540</pre>
6541
6542<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6543<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6544
6545<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6546<p>The argument and return value are floating point numbers of the same
6547 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6548
6549<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6550<p>This function returns the cosine of the specified operand, returning the same
6551 values as the libm <tt>cos</tt> functions would, and handles error conditions
6552 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6553
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6554</div>
6555
6556<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6557<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6558 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6559</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6560
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6561<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6562
6563<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6564<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
6565 floating point or vector of floating point type. Not all targets support all
6566 types however.</p>
6567
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6568<pre>
6569 declare float @llvm.pow.f32(float %Val, float %Power)
6570 declare double @llvm.pow.f64(double %Val, double %Power)
6571 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
6572 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
6573 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
6574</pre>
6575
6576<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6577<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
6578 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6579
6580<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6581<p>The second argument is a floating point power, and the first is a value to
6582 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6583
6584<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6585<p>This function returns the first value raised to the second power, returning
6586 the same values as the libm <tt>pow</tt> functions would, and handles error
6587 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6588
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]6589</div>
6590
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6591</div>
6592
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]6593<!-- _______________________________________________________________________ -->
6594<h4>
6595 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
6596</h4>
6597
6598<div>
6599
6600<h5>Syntax:</h5>
6601<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
6602 floating point or vector of floating point type. Not all targets support all
6603 types however.</p>
6604
6605<pre>
6606 declare float @llvm.exp.f32(float %Val)
6607 declare double @llvm.exp.f64(double %Val)
6608 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
6609 declare fp128 @llvm.exp.f128(fp128 %Val)
6610 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
6611</pre>
6612
6613<h5>Overview:</h5>
6614<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
6615
6616<h5>Arguments:</h5>
6617<p>The argument and return value are floating point numbers of the same
6618 type.</p>
6619
6620<h5>Semantics:</h5>
6621<p>This function returns the same values as the libm <tt>exp</tt> functions
6622 would, and handles error conditions in the same way.</p>
6623
6624</div>
6625
6626<!-- _______________________________________________________________________ -->
6627<h4>
6628 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
6629</h4>
6630
6631<div>
6632
6633<h5>Syntax:</h5>
6634<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
6635 floating point or vector of floating point type. Not all targets support all
6636 types however.</p>
6637
6638<pre>
6639 declare float @llvm.log.f32(float %Val)
6640 declare double @llvm.log.f64(double %Val)
6641 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
6642 declare fp128 @llvm.log.f128(fp128 %Val)
6643 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
6644</pre>
6645
6646<h5>Overview:</h5>
6647<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
6648
6649<h5>Arguments:</h5>
6650<p>The argument and return value are floating point numbers of the same
6651 type.</p>
6652
6653<h5>Semantics:</h5>
6654<p>This function returns the same values as the libm <tt>log</tt> functions
6655 would, and handles error conditions in the same way.</p>
6656
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]6657<h4>
6658 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
6659</h4>
6660
6661<div>
6662
6663<h5>Syntax:</h5>
6664<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
6665 floating point or vector of floating point type. Not all targets support all
6666 types however.</p>
6667
6668<pre>
6669 declare float @llvm.fma.f32(float %a, float %b, float %c)
6670 declare double @llvm.fma.f64(double %a, double %b, double %c)
6671 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
6672 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
6673 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
6674</pre>
6675
6676<h5>Overview:</h5>
Cameron Zwarichabc43e62011-07-08 22:13:55 +0000[diff] [blame]6677<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]6678 operation.</p>
6679
6680<h5>Arguments:</h5>
6681<p>The argument and return value are floating point numbers of the same
6682 type.</p>
6683
6684<h5>Semantics:</h5>
6685<p>This function returns the same values as the libm <tt>fma</tt> functions
6686 would.</p>
6687
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]6688</div>
6689
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6690<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6691<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6692 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6693</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6694
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6695<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6696
6697<p>LLVM provides intrinsics for a few important bit manipulation operations.
6698 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6699
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6700<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6701<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6702 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6703</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6704
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6705<div>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6706
6707<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6708<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6709 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
6710
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6711<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6712 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
6713 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
6714 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6715</pre>
6716
6717<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6718<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
6719 values with an even number of bytes (positive multiple of 16 bits). These
6720 are useful for performing operations on data that is not in the target's
6721 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6722
6723<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6724<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
6725 and low byte of the input i16 swapped. Similarly,
6726 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
6727 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
6728 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
6729 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
6730 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
6731 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]6732
6733</div>
6734
6735<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6736<h4>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]6737 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6738</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6739
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6740<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6741
6742<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6743<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]6744 width, or on any vector with integer elements. Not all targets support all
6745 bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6746
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6747<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6748 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6749 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6750 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6751 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
6752 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]6753 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]6754</pre>
6755
6756<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6757<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
6758 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6759
6760<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6761<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]6762 integer type, or a vector with integer elements.
6763 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6764
6765<h5>Semantics:</h5>
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]6766<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
6767 element of a vector.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6768
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6769</div>
6770
6771<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6772<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]6773 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6774</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6775
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6776<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6777
6778<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6779<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]6780 integer bit width, or any vector whose elements are integers. Not all
6781 targets support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6782
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6783<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6784 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
6785 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6786 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6787 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
6788 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]6789 declare &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src;gt)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6790</pre>
6791
6792<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6793<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
6794 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6795
6796<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6797<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]6798 integer type, or any vector type with integer element type.
6799 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6800
6801<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6802<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]6803 zeros in a variable, or within each element of the vector if the operation
6804 is of vector type. If the src == 0 then the result is the size in bits of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6805 the type of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6806
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]6807</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]6808
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6809<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6810<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]6811 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6812</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6813
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6814<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6815
6816<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6817<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]6818 integer bit width, or any vector of integer elements. Not all targets
6819 support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6820
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6821<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6822 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
6823 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6824 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]6825 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
6826 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]6827 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6828</pre>
6829
6830<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6831<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
6832 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6833
6834<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6835<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]6836 integer type, or a vectory with integer element type.. The return type
6837 must match the argument type.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6838
6839<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6840<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]6841 zeros in a variable, or within each element of a vector.
6842 If the src == 0 then the result is the size in bits of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6843 the type of src. For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6844
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]6845</div>
6846
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6847</div>
6848
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6849<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6850<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6851 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6852</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6853
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6854<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6855
6856<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]6857
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6858<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6859<h4>
6860 <a name="int_sadd_overflow">
6861 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
6862 </a>
6863</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6864
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6865<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6866
6867<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6868<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]6869 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6870
6871<pre>
6872 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
6873 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6874 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
6875</pre>
6876
6877<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6878<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6879 a signed addition of the two arguments, and indicate whether an overflow
6880 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6881
6882<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6883<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]6884 be of integer types of any bit width, but they must have the same bit
6885 width. The second element of the result structure must be of
6886 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6887 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6888
6889<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6890<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6891 a signed addition of the two variables. They return a structure &mdash; the
6892 first element of which is the signed summation, and the second element of
6893 which is a bit specifying if the signed summation resulted in an
6894 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6895
6896<h5>Examples:</h5>
6897<pre>
6898 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
6899 %sum = extractvalue {i32, i1} %res, 0
6900 %obit = extractvalue {i32, i1} %res, 1
6901 br i1 %obit, label %overflow, label %normal
6902</pre>
6903
6904</div>
6905
6906<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6907<h4>
6908 <a name="int_uadd_overflow">
6909 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
6910 </a>
6911</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6912
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6913<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6914
6915<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6916<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]6917 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6918
6919<pre>
6920 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
6921 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6922 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
6923</pre>
6924
6925<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6926<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6927 an unsigned addition of the two arguments, and indicate whether a carry
6928 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6929
6930<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6931<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]6932 be of integer types of any bit width, but they must have the same bit
6933 width. The second element of the result structure must be of
6934 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6935 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6936
6937<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6938<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6939 an unsigned addition of the two arguments. They return a structure &mdash;
6940 the first element of which is the sum, and the second element of which is a
6941 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6942
6943<h5>Examples:</h5>
6944<pre>
6945 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
6946 %sum = extractvalue {i32, i1} %res, 0
6947 %obit = extractvalue {i32, i1} %res, 1
6948 br i1 %obit, label %carry, label %normal
6949</pre>
6950
6951</div>
6952
6953<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6954<h4>
6955 <a name="int_ssub_overflow">
6956 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
6957 </a>
6958</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6959
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6960<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6961
6962<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6963<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]6964 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6965
6966<pre>
6967 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
6968 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6969 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
6970</pre>
6971
6972<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6973<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6974 a signed subtraction of the two arguments, and indicate whether an overflow
6975 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6976
6977<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6978<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]6979 be of integer types of any bit width, but they must have the same bit
6980 width. The second element of the result structure must be of
6981 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
6982 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6983
6984<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6985<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6986 a signed subtraction of the two arguments. They return a structure &mdash;
6987 the first element of which is the subtraction, and the second element of
6988 which is a bit specifying if the signed subtraction resulted in an
6989 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]6990
6991<h5>Examples:</h5>
6992<pre>
6993 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
6994 %sum = extractvalue {i32, i1} %res, 0
6995 %obit = extractvalue {i32, i1} %res, 1
6996 br i1 %obit, label %overflow, label %normal
6997</pre>
6998
6999</div>
7000
7001<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7002<h4>
7003 <a name="int_usub_overflow">
7004 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7005 </a>
7006</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7007
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7008<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7009
7010<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7011<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]7012 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7013
7014<pre>
7015 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7016 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7017 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7018</pre>
7019
7020<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7021<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7022 an unsigned subtraction of the two arguments, and indicate whether an
7023 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7024
7025<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7026<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]7027 be of integer types of any bit width, but they must have the same bit
7028 width. The second element of the result structure must be of
7029 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7030 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7031
7032<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7033<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7034 an unsigned subtraction of the two arguments. They return a structure &mdash;
7035 the first element of which is the subtraction, and the second element of
7036 which is a bit specifying if the unsigned subtraction resulted in an
7037 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7038
7039<h5>Examples:</h5>
7040<pre>
7041 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7042 %sum = extractvalue {i32, i1} %res, 0
7043 %obit = extractvalue {i32, i1} %res, 1
7044 br i1 %obit, label %overflow, label %normal
7045</pre>
7046
7047</div>
7048
7049<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7050<h4>
7051 <a name="int_smul_overflow">
7052 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7053 </a>
7054</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7055
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7056<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7057
7058<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7059<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]7060 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7061
7062<pre>
7063 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7064 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7065 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7066</pre>
7067
7068<h5>Overview:</h5>
7069
7070<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7071 a signed multiplication of the two arguments, and indicate whether an
7072 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7073
7074<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7075<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]7076 be of integer types of any bit width, but they must have the same bit
7077 width. The second element of the result structure must be of
7078 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7079 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7080
7081<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7082<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7083 a signed multiplication of the two arguments. They return a structure &mdash;
7084 the first element of which is the multiplication, and the second element of
7085 which is a bit specifying if the signed multiplication resulted in an
7086 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7087
7088<h5>Examples:</h5>
7089<pre>
7090 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7091 %sum = extractvalue {i32, i1} %res, 0
7092 %obit = extractvalue {i32, i1} %res, 1
7093 br i1 %obit, label %overflow, label %normal
7094</pre>
7095
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]7096</div>
7097
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7098<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7099<h4>
7100 <a name="int_umul_overflow">
7101 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7102 </a>
7103</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7104
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7105<div>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7106
7107<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7108<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]7109 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7110
7111<pre>
7112 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7113 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7114 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7115</pre>
7116
7117<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7118<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7119 a unsigned multiplication of the two arguments, and indicate whether an
7120 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7121
7122<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7123<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]7124 be of integer types of any bit width, but they must have the same bit
7125 width. The second element of the result structure must be of
7126 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7127 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7128
7129<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7130<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7131 an unsigned multiplication of the two arguments. They return a structure
7132 &mdash; the first element of which is the multiplication, and the second
7133 element of which is a bit specifying if the unsigned multiplication resulted
7134 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7135
7136<h5>Examples:</h5>
7137<pre>
7138 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7139 %sum = extractvalue {i32, i1} %res, 0
7140 %obit = extractvalue {i32, i1} %res, 1
7141 br i1 %obit, label %overflow, label %normal
7142</pre>
7143
7144</div>
7145
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7146</div>
7147
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7148<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7149<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7150 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7151</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7152
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7153<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7154
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7155<p>Half precision floating point is a storage-only format. This means that it is
7156 a dense encoding (in memory) but does not support computation in the
7157 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7158
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7159<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7160 value as an i16, then convert it to float with <a
7161 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7162 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7163 double etc). To store the value back to memory, it is first converted to
7164 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7165 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7166 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7167
7168<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7169<h4>
7170 <a name="int_convert_to_fp16">
7171 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7172 </a>
7173</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7174
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7175<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7176
7177<h5>Syntax:</h5>
7178<pre>
7179 declare i16 @llvm.convert.to.fp16(f32 %a)
7180</pre>
7181
7182<h5>Overview:</h5>
7183<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7184 a conversion from single precision floating point format to half precision
7185 floating point format.</p>
7186
7187<h5>Arguments:</h5>
7188<p>The intrinsic function contains single argument - the value to be
7189 converted.</p>
7190
7191<h5>Semantics:</h5>
7192<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7193 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7194 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7195 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7196
7197<h5>Examples:</h5>
7198<pre>
7199 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7200 store i16 %res, i16* @x, align 2
7201</pre>
7202
7203</div>
7204
7205<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7206<h4>
7207 <a name="int_convert_from_fp16">
7208 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7209 </a>
7210</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7211
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7212<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7213
7214<h5>Syntax:</h5>
7215<pre>
7216 declare f32 @llvm.convert.from.fp16(i16 %a)
7217</pre>
7218
7219<h5>Overview:</h5>
7220<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7221 a conversion from half precision floating point format to single precision
7222 floating point format.</p>
7223
7224<h5>Arguments:</h5>
7225<p>The intrinsic function contains single argument - the value to be
7226 converted.</p>
7227
7228<h5>Semantics:</h5>
7229<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7230 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7231 precision floating point format. The input half-float value is represented by
7232 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7233
7234<h5>Examples:</h5>
7235<pre>
7236 %a = load i16* @x, align 2
7237 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7238</pre>
7239
7240</div>
7241
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7242</div>
7243
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7244<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7245<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7246 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7247</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7248
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7249<div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7250
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7251<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7252 prefix), are described in
7253 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7254 Level Debugging</a> document.</p>
7255
7256</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7257
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7258<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7259<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7260 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7261</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7262
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7263<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7264
7265<p>The LLVM exception handling intrinsics (which all start with
7266 <tt>llvm.eh.</tt> prefix), are described in
7267 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7268 Handling</a> document.</p>
7269
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7270</div>
7271
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7272<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7273<h3>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]7274 <a name="int_trampoline">Trampoline Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7275</h3>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7276
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7277<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7278
7279<p>This intrinsic makes it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]7280 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7281 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7282 function pointer lacking the nest parameter - the caller does not need to
7283 provide a value for it. Instead, the value to use is stored in advance in a
7284 "trampoline", a block of memory usually allocated on the stack, which also
7285 contains code to splice the nest value into the argument list. This is used
7286 to implement the GCC nested function address extension.</p>
7287
7288<p>For example, if the function is
7289 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7290 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7291 follows:</p>
7292
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]7293<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]7294 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7295 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7296 %p = call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval)
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]7297 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7298</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7299
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7300<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7301 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7302
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7303<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7304<h4>
7305 <a name="int_it">
7306 '<tt>llvm.init.trampoline</tt>' Intrinsic
7307 </a>
7308</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7309
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7310<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7311
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7312<h5>Syntax:</h5>
7313<pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7314 declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7315</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7316
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7317<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7318<p>This fills the memory pointed to by <tt>tramp</tt> with code and returns a
7319 function pointer suitable for executing it.</p>
7320
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7321<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7322<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7323 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7324 sufficiently aligned block of memory; this memory is written to by the
7325 intrinsic. Note that the size and the alignment are target-specific - LLVM
7326 currently provides no portable way of determining them, so a front-end that
7327 generates this intrinsic needs to have some target-specific knowledge.
7328 The <tt>func</tt> argument must hold a function bitcast to
7329 an <tt>i8*</tt>.</p>
7330
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7331<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7332<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
7333 dependent code, turning it into a function. A pointer to this function is
7334 returned, but needs to be bitcast to an <a href="#int_trampoline">appropriate
7335 function pointer type</a> before being called. The new function's signature
7336 is the same as that of <tt>func</tt> with any arguments marked with
7337 the <tt>nest</tt> attribute removed. At most one such <tt>nest</tt> argument
7338 is allowed, and it must be of pointer type. Calling the new function is
7339 equivalent to calling <tt>func</tt> with the same argument list, but
7340 with <tt>nval</tt> used for the missing <tt>nest</tt> argument. If, after
7341 calling <tt>llvm.init.trampoline</tt>, the memory pointed to
7342 by <tt>tramp</tt> is modified, then the effect of any later call to the
7343 returned function pointer is undefined.</p>
7344
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7345</div>
7346
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7347</div>
7348
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7349<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7350<h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7351 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7352</h3>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7353
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7354<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7355
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7356<p>These intrinsic functions expand the "universal IR" of LLVM to represent
7357 hardware constructs for atomic operations and memory synchronization. This
7358 provides an interface to the hardware, not an interface to the programmer. It
7359 is aimed at a low enough level to allow any programming models or APIs
7360 (Application Programming Interfaces) which need atomic behaviors to map
7361 cleanly onto it. It is also modeled primarily on hardware behavior. Just as
7362 hardware provides a "universal IR" for source languages, it also provides a
7363 starting point for developing a "universal" atomic operation and
7364 synchronization IR.</p>
7365
7366<p>These do <em>not</em> form an API such as high-level threading libraries,
7367 software transaction memory systems, atomic primitives, and intrinsic
7368 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
7369 application libraries. The hardware interface provided by LLVM should allow
7370 a clean implementation of all of these APIs and parallel programming models.
7371 No one model or paradigm should be selected above others unless the hardware
7372 itself ubiquitously does so.</p>
7373
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7374<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7375<h4>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7376 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7377</h4>
7378
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7379<div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7380<h5>Syntax:</h5>
7381<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7382 declare void @llvm.memory.barrier(i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;, i1 &lt;device&gt;)
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7383</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7384
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7385<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7386<p>The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
7387 specific pairs of memory access types.</p>
7388
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7389<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7390<p>The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
7391 The first four arguments enables a specific barrier as listed below. The
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7392 fifth argument specifies that the barrier applies to io or device or uncached
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7393 memory.</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7394
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7395<ul>
7396 <li><tt>ll</tt>: load-load barrier</li>
7397 <li><tt>ls</tt>: load-store barrier</li>
7398 <li><tt>sl</tt>: store-load barrier</li>
7399 <li><tt>ss</tt>: store-store barrier</li>
7400 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
7401</ul>
7402
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7403<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7404<p>This intrinsic causes the system to enforce some ordering constraints upon
7405 the loads and stores of the program. This barrier does not
7406 indicate <em>when</em> any events will occur, it only enforces
7407 an <em>order</em> in which they occur. For any of the specified pairs of load
7408 and store operations (f.ex. load-load, or store-load), all of the first
7409 operations preceding the barrier will complete before any of the second
7410 operations succeeding the barrier begin. Specifically the semantics for each
7411 pairing is as follows:</p>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7412
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7413<ul>
7414 <li><tt>ll</tt>: All loads before the barrier must complete before any load
7415 after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7416 <li><tt>ls</tt>: All loads before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7417 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7418 <li><tt>ss</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7419 store after the barrier begins.</li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7420 <li><tt>sl</tt>: All stores before the barrier must complete before any
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7421 load after the barrier begins.</li>
7422</ul>
7423
7424<p>These semantics are applied with a logical "and" behavior when more than one
7425 is enabled in a single memory barrier intrinsic.</p>
7426
7427<p>Backends may implement stronger barriers than those requested when they do
7428 not support as fine grained a barrier as requested. Some architectures do
7429 not need all types of barriers and on such architectures, these become
7430 noops.</p>
7431
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7432<h5>Example:</h5>
7433<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7434%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7435%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7436 store i32 4, %ptr
7437
7438%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
Evan Cheng0b0669a2011-06-29 17:14:00 +0000[diff] [blame]7439 call void @llvm.memory.barrier(i1 false, i1 true, i1 false, i1 false, i1 true)
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7440 <i>; guarantee the above finishes</i>
7441 store i32 8, %ptr <i>; before this begins</i>
7442</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7443
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7444</div>
7445
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7446<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7447<h4>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7448 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7449</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7450
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7451<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7452
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7453<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7454<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
7455 any integer bit width and for different address spaces. Not all targets
7456 support all bit widths however.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7457
7458<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7459 declare i8 @llvm.atomic.cmp.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt;)
7460 declare i16 @llvm.atomic.cmp.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt;)
7461 declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt;)
7462 declare i64 @llvm.atomic.cmp.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7463</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7464
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7465<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7466<p>This loads a value in memory and compares it to a given value. If they are
7467 equal, it stores a new value into the memory.</p>
7468
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7469<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7470<p>The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result
7471 as well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
7472 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
7473 this integer type. While any bit width integer may be used, targets may only
7474 lower representations they support in hardware.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7475
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7476<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7477<p>This entire intrinsic must be executed atomically. It first loads the value
7478 in memory pointed to by <tt>ptr</tt> and compares it with the
7479 value <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the
7480 memory. The loaded value is yielded in all cases. This provides the
7481 equivalent of an atomic compare-and-swap operation within the SSA
7482 framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7483
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7484<h5>Examples:</h5>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7485<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7486%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7487%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7488 store i32 4, %ptr
7489
7490%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7491%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 4, %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7492 <i>; yields {i32}:result1 = 4</i>
7493%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7494%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7495
7496%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7497%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %ptr, i32 5, %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7498 <i>; yields {i32}:result2 = 8</i>
7499%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
7500
7501%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
7502</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7503
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7504</div>
7505
7506<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7507<h4>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7508 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7509</h4>
7510
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7511<div>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7512<h5>Syntax:</h5>
7513
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7514<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
7515 integer bit width. Not all targets support all bit widths however.</p>
7516
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7517<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7518 declare i8 @llvm.atomic.swap.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;val&gt;)
7519 declare i16 @llvm.atomic.swap.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;val&gt;)
7520 declare i32 @llvm.atomic.swap.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;val&gt;)
7521 declare i64 @llvm.atomic.swap.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;val&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7522</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7523
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7524<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7525<p>This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
7526 the value from memory. It then stores the value in <tt>val</tt> in the memory
7527 at <tt>ptr</tt>.</p>
7528
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7529<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7530<p>The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both
7531 the <tt>val</tt> argument and the result must be integers of the same bit
7532 width. The first argument, <tt>ptr</tt>, must be a pointer to a value of this
7533 integer type. The targets may only lower integer representations they
7534 support.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7535
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7536<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7537<p>This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
7538 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
7539 equivalent of an atomic swap operation within the SSA framework.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7540
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7541<h5>Examples:</h5>
7542<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7543%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7544%ptr = bitcast i8* %mallocP to i32*
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7545 store i32 4, %ptr
7546
7547%val1 = add i32 4, 4
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7548%result1 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val1)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7549 <i>; yields {i32}:result1 = 4</i>
7550%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
7551%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
7552
7553%val2 = add i32 1, 1
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7554%result2 = call i32 @llvm.atomic.swap.i32.p0i32(i32* %ptr, i32 %val2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7555 <i>; yields {i32}:result2 = 8</i>
7556
7557%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
7558%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
7559</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7560
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7561</div>
7562
7563<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7564<h4>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7565 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7566</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7567
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7568<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7569
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7570<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7571<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on
7572 any integer bit width. Not all targets support all bit widths however.</p>
7573
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7574<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7575 declare i8 @llvm.atomic.load.add.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7576 declare i16 @llvm.atomic.load.add.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7577 declare i32 @llvm.atomic.load.add.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7578 declare i64 @llvm.atomic.load.add.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7579</pre>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7580
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7581<h5>Overview:</h5>
7582<p>This intrinsic adds <tt>delta</tt> to the value stored in memory
7583 at <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7584
7585<h5>Arguments:</h5>
7586<p>The intrinsic takes two arguments, the first a pointer to an integer value
7587 and the second an integer value. The result is also an integer value. These
7588 integer types can have any bit width, but they must all have the same bit
7589 width. The targets may only lower integer representations they support.</p>
7590
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7591<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7592<p>This intrinsic does a series of operations atomically. It first loads the
7593 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
7594 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.</p>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7595
7596<h5>Examples:</h5>
7597<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7598%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7599%ptr = bitcast i8* %mallocP to i32*
7600 store i32 4, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7601%result1 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 4)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7602 <i>; yields {i32}:result1 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7603%result2 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 2)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7604 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7605%result3 = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %ptr, i32 5)
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7606 <i>; yields {i32}:result3 = 10</i>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7607%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7608</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7609
Andrew Lenharthab0b9492008-02-21 06:45:13 +0000[diff] [blame]7610</div>
7611
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7612<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7613<h4>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7614 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7615</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7616
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7617<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7618
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7619<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7620<p>This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
7621 any integer bit width and for different address spaces. Not all targets
7622 support all bit widths however.</p>
7623
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7624<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7625 declare i8 @llvm.atomic.load.sub.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7626 declare i16 @llvm.atomic.load.sub.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7627 declare i32 @llvm.atomic.load.sub.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7628 declare i64 @llvm.atomic.load.sub.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7629</pre>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7630
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7631<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7632<p>This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7633 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.</p>
7634
7635<h5>Arguments:</h5>
7636<p>The intrinsic takes two arguments, the first a pointer to an integer value
7637 and the second an integer value. The result is also an integer value. These
7638 integer types can have any bit width, but they must all have the same bit
7639 width. The targets may only lower integer representations they support.</p>
7640
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7641<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7642<p>This intrinsic does a series of operations atomically. It first loads the
7643 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
7644 result to <tt>ptr</tt>. It yields the original value stored
7645 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7646
7647<h5>Examples:</h5>
7648<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7649%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7650%ptr = bitcast i8* %mallocP to i32*
7651 store i32 8, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7652%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 4)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7653 <i>; yields {i32}:result1 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7654%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 2)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7655 <i>; yields {i32}:result2 = 4</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7656%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %ptr, i32 5)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7657 <i>; yields {i32}:result3 = 2</i>
7658%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
7659</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7660
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7661</div>
7662
7663<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7664<h4>
7665 <a name="int_atomic_load_and">
7666 '<tt>llvm.atomic.load.and.*</tt>' Intrinsic
7667 </a>
7668 <br>
7669 <a name="int_atomic_load_nand">
7670 '<tt>llvm.atomic.load.nand.*</tt>' Intrinsic
7671 </a>
7672 <br>
7673 <a name="int_atomic_load_or">
7674 '<tt>llvm.atomic.load.or.*</tt>' Intrinsic
7675 </a>
7676 <br>
7677 <a name="int_atomic_load_xor">
7678 '<tt>llvm.atomic.load.xor.*</tt>' Intrinsic
7679 </a>
7680</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7681
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7682<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7683
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7684<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7685<p>These are overloaded intrinsics. You can
7686 use <tt>llvm.atomic.load_and</tt>, <tt>llvm.atomic.load_nand</tt>,
7687 <tt>llvm.atomic.load_or</tt>, and <tt>llvm.atomic.load_xor</tt> on any integer
7688 bit width and for different address spaces. Not all targets support all bit
7689 widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7690
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7691<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7692 declare i8 @llvm.atomic.load.and.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7693 declare i16 @llvm.atomic.load.and.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7694 declare i32 @llvm.atomic.load.and.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7695 declare i64 @llvm.atomic.load.and.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7696</pre>
7697
7698<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7699 declare i8 @llvm.atomic.load.or.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7700 declare i16 @llvm.atomic.load.or.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7701 declare i32 @llvm.atomic.load.or.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7702 declare i64 @llvm.atomic.load.or.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7703</pre>
7704
7705<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7706 declare i8 @llvm.atomic.load.nand.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7707 declare i16 @llvm.atomic.load.nand.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7708 declare i32 @llvm.atomic.load.nand.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7709 declare i64 @llvm.atomic.load.nand.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7710</pre>
7711
7712<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7713 declare i8 @llvm.atomic.load.xor.i8.p0i32(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7714 declare i16 @llvm.atomic.load.xor.i16.p0i32(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7715 declare i32 @llvm.atomic.load.xor.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7716 declare i64 @llvm.atomic.load.xor.i64.p0i32(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7717</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7718
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7719<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7720<p>These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
7721 the value stored in memory at <tt>ptr</tt>. It yields the original value
7722 at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7723
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7724<h5>Arguments:</h5>
7725<p>These intrinsics take two arguments, the first a pointer to an integer value
7726 and the second an integer value. The result is also an integer value. These
7727 integer types can have any bit width, but they must all have the same bit
7728 width. The targets may only lower integer representations they support.</p>
7729
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7730<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7731<p>These intrinsics does a series of operations atomically. They first load the
7732 value stored at <tt>ptr</tt>. They then do the bitwise
7733 operation <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the
7734 original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7735
7736<h5>Examples:</h5>
7737<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7738%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7739%ptr = bitcast i8* %mallocP to i32*
7740 store i32 0x0F0F, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7741%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7742 <i>; yields {i32}:result0 = 0x0F0F</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7743%result1 = call i32 @llvm.atomic.load.and.i32.p0i32(i32* %ptr, i32 0xFF)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7744 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7745%result2 = call i32 @llvm.atomic.load.or.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7746 <i>; yields {i32}:result2 = 0xF0</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7747%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32(i32* %ptr, i32 0F)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7748 <i>; yields {i32}:result3 = FF</i>
7749%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
7750</pre>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7751
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7752</div>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7753
7754<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7755<h4>
7756 <a name="int_atomic_load_max">
7757 '<tt>llvm.atomic.load.max.*</tt>' Intrinsic
7758 </a>
7759 <br>
7760 <a name="int_atomic_load_min">
7761 '<tt>llvm.atomic.load.min.*</tt>' Intrinsic
7762 </a>
7763 <br>
7764 <a name="int_atomic_load_umax">
7765 '<tt>llvm.atomic.load.umax.*</tt>' Intrinsic
7766 </a>
7767 <br>
7768 <a name="int_atomic_load_umin">
7769 '<tt>llvm.atomic.load.umin.*</tt>' Intrinsic
7770 </a>
7771</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7772
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7773<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7774
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7775<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7776<p>These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
7777 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
7778 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
7779 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7780
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7781<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7782 declare i8 @llvm.atomic.load.max.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7783 declare i16 @llvm.atomic.load.max.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7784 declare i32 @llvm.atomic.load.max.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7785 declare i64 @llvm.atomic.load.max.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7786</pre>
7787
7788<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7789 declare i8 @llvm.atomic.load.min.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7790 declare i16 @llvm.atomic.load.min.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7791 declare i32 @llvm.atomic.load.min.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7792 declare i64 @llvm.atomic.load.min.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7793</pre>
7794
7795<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7796 declare i8 @llvm.atomic.load.umax.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7797 declare i16 @llvm.atomic.load.umax.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7798 declare i32 @llvm.atomic.load.umax.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7799 declare i64 @llvm.atomic.load.umax.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7800</pre>
7801
7802<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7803 declare i8 @llvm.atomic.load.umin.i8.p0i8(i8* &lt;ptr&gt;, i8 &lt;delta&gt;)
7804 declare i16 @llvm.atomic.load.umin.i16.p0i16(i16* &lt;ptr&gt;, i16 &lt;delta&gt;)
7805 declare i32 @llvm.atomic.load.umin.i32.p0i32(i32* &lt;ptr&gt;, i32 &lt;delta&gt;)
7806 declare i64 @llvm.atomic.load.umin.i64.p0i64(i64* &lt;ptr&gt;, i64 &lt;delta&gt;)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7807</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7808
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7809<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]7810<p>These intrinsics takes the signed or unsigned minimum or maximum of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7811 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
7812 original value at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7813
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7814<h5>Arguments:</h5>
7815<p>These intrinsics take two arguments, the first a pointer to an integer value
7816 and the second an integer value. The result is also an integer value. These
7817 integer types can have any bit width, but they must all have the same bit
7818 width. The targets may only lower integer representations they support.</p>
7819
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7820<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7821<p>These intrinsics does a series of operations atomically. They first load the
7822 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or
7823 max <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They
7824 yield the original value stored at <tt>ptr</tt>.</p>
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7825
7826<h5>Examples:</h5>
7827<pre>
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]7828%mallocP = tail call i8* @malloc(i32 ptrtoint (i32* getelementptr (i32* null, i32 1) to i32))
7829%ptr = bitcast i8* %mallocP to i32*
7830 store i32 7, %ptr
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7831%result0 = call i32 @llvm.atomic.load.min.i32.p0i32(i32* %ptr, i32 -2)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7832 <i>; yields {i32}:result0 = 7</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7833%result1 = call i32 @llvm.atomic.load.max.i32.p0i32(i32* %ptr, i32 8)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7834 <i>; yields {i32}:result1 = -2</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7835%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32(i32* %ptr, i32 10)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7836 <i>; yields {i32}:result2 = 8</i>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7837%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32(i32* %ptr, i32 30)
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7838 <i>; yields {i32}:result3 = 8</i>
7839%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
7840</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7841
Mon P Wang28873102008-06-25 08:15:39 +0000[diff] [blame]7842</div>
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7843
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7844</div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7845
7846<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7847<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7848 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7849</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7850
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7851<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7852
7853<p>This class of intrinsics exists to information about the lifetime of memory
7854 objects and ranges where variables are immutable.</p>
7855
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7856<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7857<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7858 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7859</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7860
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7861<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7862
7863<h5>Syntax:</h5>
7864<pre>
7865 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7866</pre>
7867
7868<h5>Overview:</h5>
7869<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7870 object's lifetime.</p>
7871
7872<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7873<p>The first argument is a constant integer representing the size of the
7874 object, or -1 if it is variable sized. The second argument is a pointer to
7875 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7876
7877<h5>Semantics:</h5>
7878<p>This intrinsic indicates that before this point in the code, the value of the
7879 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]7880 never be used and has an undefined value. A load from the pointer that
7881 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7882 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7883
7884</div>
7885
7886<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7887<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7888 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7889</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7890
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7891<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7892
7893<h5>Syntax:</h5>
7894<pre>
7895 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7896</pre>
7897
7898<h5>Overview:</h5>
7899<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
7900 object's lifetime.</p>
7901
7902<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7903<p>The first argument is a constant integer representing the size of the
7904 object, or -1 if it is variable sized. The second argument is a pointer to
7905 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7906
7907<h5>Semantics:</h5>
7908<p>This intrinsic indicates that after this point in the code, the value of the
7909 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
7910 never be used and has an undefined value. Any stores into the memory object
7911 following this intrinsic may be removed as dead.
7912
7913</div>
7914
7915<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7916<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7917 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7918</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7919
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7920<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7921
7922<h5>Syntax:</h5>
7923<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +0000[diff] [blame]7924 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7925</pre>
7926
7927<h5>Overview:</h5>
7928<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
7929 a memory object will not change.</p>
7930
7931<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7932<p>The first argument is a constant integer representing the size of the
7933 object, or -1 if it is variable sized. The second argument is a pointer to
7934 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7935
7936<h5>Semantics:</h5>
7937<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
7938 the return value, the referenced memory location is constant and
7939 unchanging.</p>
7940
7941</div>
7942
7943<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7944<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7945 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7946</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7947
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7948<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7949
7950<h5>Syntax:</h5>
7951<pre>
7952 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7953</pre>
7954
7955<h5>Overview:</h5>
7956<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
7957 a memory object are mutable.</p>
7958
7959<h5>Arguments:</h5>
7960<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7961 The second argument is a constant integer representing the size of the
7962 object, or -1 if it is variable sized and the third argument is a pointer
7963 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7964
7965<h5>Semantics:</h5>
7966<p>This intrinsic indicates that the memory is mutable again.</p>
7967
7968</div>
7969
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7970</div>
7971
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]7972<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7973<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7974 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7975</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7976
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7977<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7978
7979<p>This class of intrinsics is designed to be generic and has no specific
7980 purpose.</p>
7981
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7982<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7983<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7984 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7985</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7986
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7987<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7988
7989<h5>Syntax:</h5>
7990<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7991 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]7992</pre>
7993
7994<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7995<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7996
7997<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7998<p>The first argument is a pointer to a value, the second is a pointer to a
7999 global string, the third is a pointer to a global string which is the source
8000 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8001
8002<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8003<p>This intrinsic allows annotation of local variables with arbitrary strings.
8004 This can be useful for special purpose optimizations that want to look for
8005 these annotations. These have no other defined use, they are ignored by code
8006 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8007
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8008</div>
8009
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8010<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8011<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]8012 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8013</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8014
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8015<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8016
8017<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8018<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8019 any integer bit width.</p>
8020
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8021<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8022 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8023 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8024 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8025 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8026 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]8027</pre>
8028
8029<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8030<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8031
8032<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8033<p>The first argument is an integer value (result of some expression), the
8034 second is a pointer to a global string, the third is a pointer to a global
8035 string which is the source file name, and the last argument is the line
8036 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8037
8038<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8039<p>This intrinsic allows annotations to be put on arbitrary expressions with
8040 arbitrary strings. This can be useful for special purpose optimizations that
8041 want to look for these annotations. These have no other defined use, they
8042 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8043
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8044</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8045
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8046<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8047<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8048 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8049</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8050
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8051<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8052
8053<h5>Syntax:</h5>
8054<pre>
8055 declare void @llvm.trap()
8056</pre>
8057
8058<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8059<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8060
8061<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8062<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8063
8064<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8065<p>This intrinsics is lowered to the target dependent trap instruction. If the
8066 target does not have a trap instruction, this intrinsic will be lowered to
8067 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8068
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8069</div>
8070
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8071<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8072<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +0000[diff] [blame]8073 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8074</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8075
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8076<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8077
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8078<h5>Syntax:</h5>
8079<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8080 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8081</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8082
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8083<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8084<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8085 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8086 ensure that it is placed on the stack before local variables.</p>
8087
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8088<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8089<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8090 arguments. The first argument is the value loaded from the stack
8091 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8092 that has enough space to hold the value of the guard.</p>
8093
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8094<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8095<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8096 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8097 stack. This is to ensure that if a local variable on the stack is
8098 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8099 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8100 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8101 function.</p>
8102
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8103</div>
8104
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8105<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8106<h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8107 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8108</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8109
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8110<div>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8111
8112<h5>Syntax:</h5>
8113<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8114 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8115 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8116</pre>
8117
8118<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8119<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8120 the optimizers to determine at compile time whether a) an operation (like
8121 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8122 runtime check for overflow isn't necessary. An object in this context means
8123 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8124
8125<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8126<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8127 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8128 is a boolean 0 or 1. This argument determines whether you want the
8129 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]8130 1, variables are not allowed.</p>
8131
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8132<h5>Semantics:</h5>
8133<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8134 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8135 depending on the <tt>type</tt> argument, if the size cannot be determined at
8136 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8137
8138</div>
8139
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8140</div>
8141
8142</div>
8143
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8144<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8145<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8146<address>
8147 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8148 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8149 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8150 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]8151
8152 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]8153 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8154 Last modified: $Date$
8155</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]8156
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]8157</body>
8158</html>