Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 7302d80490feabfc8a01bee0fa698aab55169544 / . / docs / LangRef.html
blob: ee8eda49bf61385f6301137af7e8ea186922b6c2 [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]8 <meta name="description"
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]14
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]15<h1>LLVM Language Reference Manual</h1>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a>
24 <ol>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]25 <li><a href="#linkage_private">'<tt>private</tt>' Linkage</a></li>
26 <li><a href="#linkage_linker_private">'<tt>linker_private</tt>' Linkage</a></li>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]27 <li><a href="#linkage_linker_private_weak">'<tt>linker_private_weak</tt>' Linkage</a></li>
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]28 <li><a href="#linkage_linker_private_weak_def_auto">'<tt>linker_private_weak_def_auto</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]29 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
30 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
31 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
32 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
33 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
34 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
35 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner5a2d8752009-10-10 18:26:06 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]37 <li><a href="#linkage_weak">'<tt>weak_odr</tt>' Linkage</a></li>
Bill Wendlingf7f06102011-10-11 06:41:28 +0000[diff] [blame]38 <li><a href="#linkage_external">'<tt>external</tt>' Linkage</a></li>
Bill Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]39 <li><a href="#linkage_dllimport">'<tt>dllimport</tt>' Linkage</a></li>
40 <li><a href="#linkage_dllexport">'<tt>dllexport</tt>' Linkage</a></li>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]41 </ol>
42 </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]43 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]44 <li><a href="#namedtypes">Named Types</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]45 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]46 <li><a href="#functionstructure">Functions</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]47 <li><a href="#aliasstructure">Aliases</a></li>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]48 <li><a href="#namedmetadatastructure">Named Metadata</a></li>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]49 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]50 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]51 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]52 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]53 <li><a href="#datalayout">Data Layout</a></li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]54 <li><a href="#pointeraliasing">Pointer Aliasing Rules</a></li>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]55 <li><a href="#volatile">Volatile Memory Accesses</a></li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]56 <li><a href="#memmodel">Memory Model for Concurrent Operations</a></li>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]57 <li><a href="#ordering">Atomic Memory Ordering Constraints</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]58 </ol>
59 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]60 <li><a href="#typesystem">Type System</a>
61 <ol>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]62 <li><a href="#t_classifications">Type Classifications</a></li>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]63 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]64 <ol>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]65 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]66 <li><a href="#t_floating">Floating Point Types</a></li>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]67 <li><a href="#t_x86mmx">X86mmx Type</a></li>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]68 <li><a href="#t_void">Void Type</a></li>
69 <li><a href="#t_label">Label Type</a></li>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]70 <li><a href="#t_metadata">Metadata Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]71 </ol>
72 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]73 <li><a href="#t_derived">Derived Types</a>
74 <ol>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]75 <li><a href="#t_aggregate">Aggregate Types</a>
76 <ol>
77 <li><a href="#t_array">Array Type</a></li>
78 <li><a href="#t_struct">Structure Type</a></li>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]79 <li><a href="#t_opaque">Opaque Structure Types</a></li>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]80 <li><a href="#t_vector">Vector Type</a></li>
81 </ol>
82 </li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]83 <li><a href="#t_function">Function Type</a></li>
84 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]85 </ol>
86 </li>
87 </ol>
88 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]89 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]90 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]91 <li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]92 <li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]93 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
94 <li><a href="#undefvalues">Undefined Values</a></li>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]95 <li><a href="#poisonvalues">Poison Values</a></li>
Chris Lattnerf9d078e2009-10-27 21:19:13 +0000[diff] [blame]96 <li><a href="#blockaddress">Addresses of Basic Blocks</a></li>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]97 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]98 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]99 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]100 <li><a href="#othervalues">Other Values</a>
101 <ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]102 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]103 <li><a href="#metadata">Metadata Nodes and Metadata Strings</a>
104 <ol>
105 <li><a href="#tbaa">'<tt>tbaa</tt>' Metadata</a></li>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]106 <li><a href="#fpaccuracy">'<tt>fpaccuracy</tt>' Metadata</a></li>
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]107 </ol>
108 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]109 </ol>
110 </li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]111 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
112 <ol>
113 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]114 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
115 Global Variable</a></li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]116 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
117 Global Variable</a></li>
118 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
119 Global Variable</a></li>
120 </ol>
121 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]122 <li><a href="#instref">Instruction Reference</a>
123 <ol>
124 <li><a href="#terminators">Terminator Instructions</a>
125 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]126 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
127 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]128 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]129 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]130 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]131 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]132 <li><a href="#i_resume">'<tt>resume</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]133 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]134 </ol>
135 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]136 <li><a href="#binaryops">Binary Operations</a>
137 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]138 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]139 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]140 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]141 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]142 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]143 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]144 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
145 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
146 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]147 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
148 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
149 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]150 </ol>
151 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]152 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
153 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]154 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
155 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
156 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]157 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]158 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]159 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]160 </ol>
161 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]162 <li><a href="#vectorops">Vector Operations</a>
163 <ol>
164 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
165 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
166 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]167 </ol>
168 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]169 <li><a href="#aggregateops">Aggregate Operations</a>
170 <ol>
171 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
172 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
173 </ol>
174 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]175 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]176 <ol>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]177 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
178 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
179 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
180 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
181 <li><a href="#i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a></li>
182 <li><a href="#i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]183 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]184 </ol>
185 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]186 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]187 <ol>
188 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
189 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
190 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
191 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
192 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]193 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
194 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
195 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
196 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]197 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
198 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]199 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]200 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]201 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]202 <li><a href="#otherops">Other Operations</a>
203 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]204 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
205 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]206 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]207 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]208 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]209 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]210 <li><a href="#i_landingpad">'<tt>landingpad</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]211 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]212 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]213 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]214 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]215 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]216 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]217 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
218 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]219 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
220 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
221 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]222 </ol>
223 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]224 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
225 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]226 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
227 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
228 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]229 </ol>
230 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]231 <li><a href="#int_codegen">Code Generator Intrinsics</a>
232 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]233 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
234 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
235 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
236 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
237 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
238 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000[diff] [blame]239 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]240 </ol>
241 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]242 <li><a href="#int_libc">Standard C Library Intrinsics</a>
243 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]244 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
245 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
246 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
247 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
248 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]249 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
250 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
251 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohman08b280b2011-05-27 00:36:31 +0000[diff] [blame]252 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
253 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]254 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]255 </ol>
256 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]257 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]258 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]259 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]260 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
261 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
262 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]263 </ol>
264 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]265 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
266 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]267 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
268 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
269 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
270 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
271 <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]272 <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]273 </ol>
274 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]275 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
276 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]277 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
278 <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]279 </ol>
280 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]281 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]282 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]283 <li><a href="#int_trampoline">Trampoline Intrinsics</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]284 <ol>
285 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]286 <li><a href="#int_at">'<tt>llvm.adjust.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]287 </ol>
288 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]289 <li><a href="#int_memorymarkers">Memory Use Markers</a>
290 <ol>
Jakub Staszak8e1b12a2011-12-04 20:44:25 +0000[diff] [blame]291 <li><a href="#int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a></li>
292 <li><a href="#int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a></li>
293 <li><a href="#int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a></li>
294 <li><a href="#int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a></li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]295 </ol>
296 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]297 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]298 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]299 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]300 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]301 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]302 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]303 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]304 '<tt>llvm.trap</tt>' Intrinsic</a></li>
305 <li><a href="#int_stackprotector">
306 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]307 <li><a href="#int_objectsize">
308 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]309 <li><a href="#int_expect">
310 '<tt>llvm.expect</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]311 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]312 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]313 </ol>
314 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]315</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]316
317<div class="doc_author">
318 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
319 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]320</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]321
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]322<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]323<h2><a name="abstract">Abstract</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]324<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]325
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]326<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]327
328<p>This document is a reference manual for the LLVM assembly language. LLVM is
329 a Static Single Assignment (SSA) based representation that provides type
330 safety, low-level operations, flexibility, and the capability of representing
331 'all' high-level languages cleanly. It is the common code representation
332 used throughout all phases of the LLVM compilation strategy.</p>
333
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]334</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]335
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]336<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]337<h2><a name="introduction">Introduction</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]338<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]339
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]340<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]341
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]342<p>The LLVM code representation is designed to be used in three different forms:
343 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
344 for fast loading by a Just-In-Time compiler), and as a human readable
345 assembly language representation. This allows LLVM to provide a powerful
346 intermediate representation for efficient compiler transformations and
347 analysis, while providing a natural means to debug and visualize the
348 transformations. The three different forms of LLVM are all equivalent. This
349 document describes the human readable representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]350
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]351<p>The LLVM representation aims to be light-weight and low-level while being
352 expressive, typed, and extensible at the same time. It aims to be a
353 "universal IR" of sorts, by being at a low enough level that high-level ideas
354 may be cleanly mapped to it (similar to how microprocessors are "universal
355 IR's", allowing many source languages to be mapped to them). By providing
356 type information, LLVM can be used as the target of optimizations: for
357 example, through pointer analysis, it can be proven that a C automatic
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]358 variable is never accessed outside of the current function, allowing it to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]359 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]360
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]361<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]362<h4>
363 <a name="wellformed">Well-Formedness</a>
364</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]365
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]366<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]367
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]368<p>It is important to note that this document describes 'well formed' LLVM
369 assembly language. There is a difference between what the parser accepts and
370 what is considered 'well formed'. For example, the following instruction is
371 syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]372
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]373<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]374%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]375</pre>
376
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]377<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
378 LLVM infrastructure provides a verification pass that may be used to verify
379 that an LLVM module is well formed. This pass is automatically run by the
380 parser after parsing input assembly and by the optimizer before it outputs
381 bitcode. The violations pointed out by the verifier pass indicate bugs in
382 transformation passes or input to the parser.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]383
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]384</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]385
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]386</div>
387
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame]388<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]389
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]390<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]391<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]392<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]393
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]394<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]395
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]396<p>LLVM identifiers come in two basic types: global and local. Global
397 identifiers (functions, global variables) begin with the <tt>'@'</tt>
398 character. Local identifiers (register names, types) begin with
399 the <tt>'%'</tt> character. Additionally, there are three different formats
400 for identifiers, for different purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]401
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]402<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]403 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]404 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
405 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
406 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
407 other characters in their names can be surrounded with quotes. Special
408 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
409 ASCII code for the character in hexadecimal. In this way, any character
410 can be used in a name value, even quotes themselves.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]411
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]412 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]413 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]414
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]415 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]416 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]417</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]418
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]419<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]420 don't need to worry about name clashes with reserved words, and the set of
421 reserved words may be expanded in the future without penalty. Additionally,
422 unnamed identifiers allow a compiler to quickly come up with a temporary
423 variable without having to avoid symbol table conflicts.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]424
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]425<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]426 languages. There are keywords for different opcodes
427 ('<tt><a href="#i_add">add</a></tt>',
428 '<tt><a href="#i_bitcast">bitcast</a></tt>',
429 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
430 ('<tt><a href="#t_void">void</a></tt>',
431 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
432 reserved words cannot conflict with variable names, because none of them
433 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]434
435<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]436 '<tt>%X</tt>' by 8:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]437
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]438<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]439
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]440<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]441%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]442</pre>
443
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]444<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]445
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]446<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]447%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]448</pre>
449
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]450<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]451
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]452<pre class="doc_code">
Gabor Greifec58f752009-10-28 13:05:07 +0000[diff] [blame]453%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
454%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]455%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]456</pre>
457
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]458<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
459 lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]460
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]461<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]462 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]463 line.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]464
465 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]466 assigned to a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]467
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]468 <li>Unnamed temporaries are numbered sequentially</li>
469</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]470
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]471<p>It also shows a convention that we follow in this document. When
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]472 demonstrating instructions, we will follow an instruction with a comment that
473 defines the type and name of value produced. Comments are shown in italic
474 text.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]475
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]476</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]477
478<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]479<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]480<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]481<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]482<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]483<h3>
484 <a name="modulestructure">Module Structure</a>
485</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]486
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]487<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]488
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]489<p>LLVM programs are composed of "Module"s, each of which is a translation unit
490 of the input programs. Each module consists of functions, global variables,
491 and symbol table entries. Modules may be combined together with the LLVM
492 linker, which merges function (and global variable) definitions, resolves
493 forward declarations, and merges symbol table entries. Here is an example of
494 the "hello world" module:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]495
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]496<pre class="doc_code">
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]497<i>; Declare the string constant as a global constant.</i>&nbsp;
Nick Lewyckydb9cd762011-01-29 01:09:53 +0000[diff] [blame]498<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]499
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]500<i>; External declaration of the puts function</i>&nbsp;
501<a href="#functionstructure">declare</a> i32 @puts(i8*) <i>; i32 (i8*)* </i>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]502
503<i>; Definition of main function</i>
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]504define i32 @main() { <i>; i32()* </i>&nbsp;
505 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
506 %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]507
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]508 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
509 <a href="#i_call">call</a> i32 @puts(i8* %cast210) <i>; i32</i>&nbsp;
510 <a href="#i_ret">ret</a> i32 0&nbsp;
511}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]512
513<i>; Named metadata</i>
514!1 = metadata !{i32 41}
515!foo = !{!1, null}
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]516</pre>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]517
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]518<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]519 "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]520 a <a href="#functionstructure">function definition</a> for
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]521 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
522 "<tt>foo"</tt>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]523
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]524<p>In general, a module is made up of a list of global values, where both
525 functions and global variables are global values. Global values are
526 represented by a pointer to a memory location (in this case, a pointer to an
527 array of char, and a pointer to a function), and have one of the
528 following <a href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]529
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]530</div>
531
532<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]533<h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]534 <a name="linkage">Linkage Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]535</h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]536
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]537<div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]538
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]539<p>All Global Variables and Functions have one of the following types of
540 linkage:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]541
542<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]543 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]544 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
545 by objects in the current module. In particular, linking code into a
546 module with an private global value may cause the private to be renamed as
547 necessary to avoid collisions. Because the symbol is private to the
548 module, all references can be updated. This doesn't show up in any symbol
549 table in the object file.</dd>
Rafael Espindolabb46f522009-01-15 20:18:42 +0000[diff] [blame]550
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]551 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]552 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
553 assembler and evaluated by the linker. Unlike normal strong symbols, they
554 are removed by the linker from the final linked image (executable or
555 dynamic library).</dd>
556
557 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
558 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
559 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
560 linker. The symbols are removed by the linker from the final linked image
561 (executable or dynamic library).</dd>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]562
Bill Wendling55ae5152010-08-20 22:05:50 +0000[diff] [blame]563 <dt><tt><b><a name="linkage_linker_private_weak_def_auto">linker_private_weak_def_auto</a></b></tt></dt>
564 <dd>Similar to "<tt>linker_private_weak</tt>", but it's known that the address
565 of the object is not taken. For instance, functions that had an inline
566 definition, but the compiler decided not to inline it. Note,
567 unlike <tt>linker_private</tt> and <tt>linker_private_weak</tt>,
568 <tt>linker_private_weak_def_auto</tt> may have only <tt>default</tt>
569 visibility. The symbols are removed by the linker from the final linked
570 image (executable or dynamic library).</dd>
571
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]572 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling07d31772010-06-29 22:34:52 +0000[diff] [blame]573 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]574 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
575 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]576
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]577 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]578 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]579 into the object file corresponding to the LLVM module. They exist to
580 allow inlining and other optimizations to take place given knowledge of
581 the definition of the global, which is known to be somewhere outside the
582 module. Globals with <tt>available_externally</tt> linkage are allowed to
583 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
584 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]585
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]586 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]587 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner873187c2010-01-09 19:15:14 +0000[diff] [blame]588 the same name when linkage occurs. This can be used to implement
589 some forms of inline functions, templates, or other code which must be
590 generated in each translation unit that uses it, but where the body may
591 be overridden with a more definitive definition later. Unreferenced
592 <tt>linkonce</tt> globals are allowed to be discarded. Note that
593 <tt>linkonce</tt> linkage does not actually allow the optimizer to
594 inline the body of this function into callers because it doesn't know if
595 this definition of the function is the definitive definition within the
596 program or whether it will be overridden by a stronger definition.
597 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
598 linkage.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]599
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]600 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]601 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
602 <tt>linkonce</tt> linkage, except that unreferenced globals with
603 <tt>weak</tt> linkage may not be discarded. This is used for globals that
604 are declared "weak" in C source code.</dd>
605
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]606 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]607 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
608 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
609 global scope.
610 Symbols with "<tt>common</tt>" linkage are merged in the same way as
611 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]612 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]613 must have a zero initializer, and may not be marked '<a
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]614 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
615 have common linkage.</dd>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]616
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]617
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]618 <dt><tt><b><a name="linkage_appending">appending</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]619 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]620 pointer to array type. When two global variables with appending linkage
621 are linked together, the two global arrays are appended together. This is
622 the LLVM, typesafe, equivalent of having the system linker append together
623 "sections" with identical names when .o files are linked.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]624
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]625 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]626 <dd>The semantics of this linkage follow the ELF object file model: the symbol
627 is weak until linked, if not linked, the symbol becomes null instead of
628 being an undefined reference.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]629
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]630 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
631 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]632 <dd>Some languages allow differing globals to be merged, such as two functions
633 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]634 that only equivalent globals are ever merged (the "one definition rule"
635 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]636 and <tt>weak_odr</tt> linkage types to indicate that the global will only
637 be merged with equivalent globals. These linkage types are otherwise the
638 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]639
Bill Wendling5c3a9f72011-11-04 20:40:41 +0000[diff] [blame]640 <dt><tt><b><a name="linkage_external">external</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]641 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]642 visible, meaning that it participates in linkage and can be used to
643 resolve external symbol references.</dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]644</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]645
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]646<p>The next two types of linkage are targeted for Microsoft Windows platform
647 only. They are designed to support importing (exporting) symbols from (to)
648 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]649
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]650<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]651 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]652 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]653 or variable via a global pointer to a pointer that is set up by the DLL
654 exporting the symbol. On Microsoft Windows targets, the pointer name is
655 formed by combining <code>__imp_</code> and the function or variable
656 name.</dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]657
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]658 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]659 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]660 pointer to a pointer in a DLL, so that it can be referenced with the
661 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
662 name is formed by combining <code>__imp_</code> and the function or
663 variable name.</dd>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]664</dl>
665
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]666<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
667 another module defined a "<tt>.LC0</tt>" variable and was linked with this
668 one, one of the two would be renamed, preventing a collision. Since
669 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
670 declarations), they are accessible outside of the current module.</p>
671
672<p>It is illegal for a function <i>declaration</i> to have any linkage type
Bill Wendlingf7f06102011-10-11 06:41:28 +0000[diff] [blame]673 other than <tt>external</tt>, <tt>dllimport</tt>
674 or <tt>extern_weak</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]675
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]676<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]677 or <tt>weak_odr</tt> linkages.</p>
678
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]679</div>
680
681<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]682<h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]683 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]684</h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]685
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]686<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]687
688<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]689 and <a href="#i_invoke">invokes</a> can all have an optional calling
690 convention specified for the call. The calling convention of any pair of
691 dynamic caller/callee must match, or the behavior of the program is
692 undefined. The following calling conventions are supported by LLVM, and more
693 may be added in the future:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]694
695<dl>
696 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]697 <dd>This calling convention (the default if no other calling convention is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]698 specified) matches the target C calling conventions. This calling
699 convention supports varargs function calls and tolerates some mismatch in
700 the declared prototype and implemented declaration of the function (as
701 does normal C).</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]702
703 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]704 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]705 (e.g. by passing things in registers). This calling convention allows the
706 target to use whatever tricks it wants to produce fast code for the
707 target, without having to conform to an externally specified ABI
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]708 (Application Binary Interface).
709 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]710 when this or the GHC convention is used.</a> This calling convention
711 does not support varargs and requires the prototype of all callees to
712 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]713
714 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]715 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]716 as possible under the assumption that the call is not commonly executed.
717 As such, these calls often preserve all registers so that the call does
718 not break any live ranges in the caller side. This calling convention
719 does not support varargs and requires the prototype of all callees to
720 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]721
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]722 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
723 <dd>This calling convention has been implemented specifically for use by the
724 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
725 It passes everything in registers, going to extremes to achieve this by
726 disabling callee save registers. This calling convention should not be
727 used lightly but only for specific situations such as an alternative to
728 the <em>register pinning</em> performance technique often used when
729 implementing functional programming languages.At the moment only X86
730 supports this convention and it has the following limitations:
731 <ul>
732 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
733 floating point types are supported.</li>
734 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
735 6 floating point parameters.</li>
736 </ul>
737 This calling convention supports
738 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
739 requires both the caller and callee are using it.
740 </dd>
741
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]742 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]743 <dd>Any calling convention may be specified by number, allowing
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]744 target-specific calling conventions to be used. Target specific calling
745 conventions start at 64.</dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]746</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]747
748<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]749 support Pascal conventions or any other well-known target-independent
750 convention.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]751
752</div>
753
754<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]755<h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]756 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]757</h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]758
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]759<div>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]760
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]761<p>All Global Variables and Functions have one of the following visibility
762 styles:</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]763
764<dl>
765 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]766 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]767 that the declaration is visible to other modules and, in shared libraries,
768 means that the declared entity may be overridden. On Darwin, default
769 visibility means that the declaration is visible to other modules. Default
770 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]771
772 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]773 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]774 object if they are in the same shared object. Usually, hidden visibility
775 indicates that the symbol will not be placed into the dynamic symbol
776 table, so no other module (executable or shared library) can reference it
777 directly.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]778
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]779 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]780 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]781 the dynamic symbol table, but that references within the defining module
782 will bind to the local symbol. That is, the symbol cannot be overridden by
783 another module.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]784</dl>
785
786</div>
787
788<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]789<h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]790 <a name="namedtypes">Named Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]791</h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]792
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]793<div>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]794
795<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]796 it easier to read the IR and make the IR more condensed (particularly when
797 recursive types are involved). An example of a name specification is:</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]798
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]799<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]800%mytype = type { %mytype*, i32 }
801</pre>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]802
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]803<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattnerdc65f222010-08-17 23:26:04 +0000[diff] [blame]804 "<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]805 is expected with the syntax "%mytype".</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]806
807<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]808 and that you can therefore specify multiple names for the same type. This
809 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
810 uses structural typing, the name is not part of the type. When printing out
811 LLVM IR, the printer will pick <em>one name</em> to render all types of a
812 particular shape. This means that if you have code where two different
813 source types end up having the same LLVM type, that the dumper will sometimes
814 print the "wrong" or unexpected type. This is an important design point and
815 isn't going to change.</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]816
817</div>
818
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]819<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]820<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]821 <a name="globalvars">Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]822</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]823
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]824<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]825
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]826<p>Global variables define regions of memory allocated at compilation time
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]827 instead of run-time. Global variables may optionally be initialized, may
828 have an explicit section to be placed in, and may have an optional explicit
829 alignment specified. A variable may be defined as "thread_local", which
830 means that it will not be shared by threads (each thread will have a
831 separated copy of the variable). A variable may be defined as a global
832 "constant," which indicates that the contents of the variable
833 will <b>never</b> be modified (enabling better optimization, allowing the
834 global data to be placed in the read-only section of an executable, etc).
835 Note that variables that need runtime initialization cannot be marked
836 "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]837
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]838<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
839 constant, even if the final definition of the global is not. This capability
840 can be used to enable slightly better optimization of the program, but
841 requires the language definition to guarantee that optimizations based on the
842 'constantness' are valid for the translation units that do not include the
843 definition.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]844
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]845<p>As SSA values, global variables define pointer values that are in scope
846 (i.e. they dominate) all basic blocks in the program. Global variables
847 always define a pointer to their "content" type because they describe a
848 region of memory, and all memory objects in LLVM are accessed through
849 pointers.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]850
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]851<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
852 that the address is not significant, only the content. Constants marked
Rafael Espindolaa5eaa862011-01-15 08:20:57 +0000[diff] [blame]853 like this can be merged with other constants if they have the same
854 initializer. Note that a constant with significant address <em>can</em>
855 be merged with a <tt>unnamed_addr</tt> constant, the result being a
856 constant whose address is significant.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]857
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]858<p>A global variable may be declared to reside in a target-specific numbered
859 address space. For targets that support them, address spaces may affect how
860 optimizations are performed and/or what target instructions are used to
861 access the variable. The default address space is zero. The address space
862 qualifier must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000[diff] [blame]863
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]864<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]865 supports it, it will emit globals to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]866
Chris Lattnerce99fa92010-04-28 00:13:42 +0000[diff] [blame]867<p>An explicit alignment may be specified for a global, which must be a power
868 of 2. If not present, or if the alignment is set to zero, the alignment of
869 the global is set by the target to whatever it feels convenient. If an
870 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner2d4b8ee2010-04-28 00:31:12 +0000[diff] [blame]871 alignment. Targets and optimizers are not allowed to over-align the global
872 if the global has an assigned section. In this case, the extra alignment
873 could be observable: for example, code could assume that the globals are
874 densely packed in their section and try to iterate over them as an array,
875 alignment padding would break this iteration.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]876
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]877<p>For example, the following defines a global in a numbered address space with
878 an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]879
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]880<pre class="doc_code">
Dan Gohman398873c2009-01-11 00:40:00 +0000[diff] [blame]881@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]882</pre>
883
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]884</div>
885
886
887<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]888<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]889 <a name="functionstructure">Functions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]890</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]891
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]892<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]893
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]894<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]895 optional <a href="#linkage">linkage type</a>, an optional
896 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]897 <a href="#callingconv">calling convention</a>,
898 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]899 <a href="#paramattrs">parameter attribute</a> for the return type, a function
900 name, a (possibly empty) argument list (each with optional
901 <a href="#paramattrs">parameter attributes</a>), optional
902 <a href="#fnattrs">function attributes</a>, an optional section, an optional
903 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
904 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]905
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]906<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
907 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]908 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]909 <a href="#callingconv">calling convention</a>,
910 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]911 <a href="#paramattrs">parameter attribute</a> for the return type, a function
912 name, a possibly empty list of arguments, an optional alignment, and an
913 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]914
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]915<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]916 (Control Flow Graph) for the function. Each basic block may optionally start
917 with a label (giving the basic block a symbol table entry), contains a list
918 of instructions, and ends with a <a href="#terminators">terminator</a>
919 instruction (such as a branch or function return).</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]920
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]921<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]922 executed on entrance to the function, and it is not allowed to have
923 predecessor basic blocks (i.e. there can not be any branches to the entry
924 block of a function). Because the block can have no predecessors, it also
925 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]926
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]927<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]928 supports it, it will emit functions to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]929
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]930<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]931 the alignment is set to zero, the alignment of the function is set by the
932 target to whatever it feels convenient. If an explicit alignment is
933 specified, the function is forced to have at least that much alignment. All
934 alignments must be a power of 2.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]935
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]936<p>If the <tt>unnamed_addr</tt> attribute is given, the address is know to not
Bill Wendling5c3a9f72011-11-04 20:40:41 +0000[diff] [blame]937 be significant and two identical functions can be merged.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]938
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]939<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]940<pre class="doc_code">
Chris Lattner50ad45c2008-10-13 16:55:18 +0000[diff] [blame]941define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]942 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
943 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
944 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
945 [<a href="#gc">gc</a>] { ... }
946</pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]947
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]948</div>
949
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]950<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]951<h3>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]952 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]953</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]954
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]955<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]956
957<p>Aliases act as "second name" for the aliasee value (which can be either
958 function, global variable, another alias or bitcast of global value). Aliases
959 may have an optional <a href="#linkage">linkage type</a>, and an
960 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]961
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]962<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]963<pre class="doc_code">
Duncan Sands0b23ac12008-09-12 20:48:21 +0000[diff] [blame]964@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]965</pre>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]966
967</div>
968
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]969<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]970<h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]971 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]972</h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]973
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]974<div>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]975
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]976<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]977 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]978 a named metadata.</p>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]979
980<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]981<pre class="doc_code">
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]982; Some unnamed metadata nodes, which are referenced by the named metadata.
983!0 = metadata !{metadata !"zero"}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]984!1 = metadata !{metadata !"one"}
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]985!2 = metadata !{metadata !"two"}
Dan Gohman1005bc52010-07-13 19:48:13 +0000[diff] [blame]986; A named metadata.
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]987!name = !{!0, !1, !2}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]988</pre>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]989
990</div>
991
992<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]993<h3>
994 <a name="paramattrs">Parameter Attributes</a>
995</h3>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]996
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]997<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]998
999<p>The return type and each parameter of a function type may have a set of
1000 <i>parameter attributes</i> associated with them. Parameter attributes are
1001 used to communicate additional information about the result or parameters of
1002 a function. Parameter attributes are considered to be part of the function,
1003 not of the function type, so functions with different parameter attributes
1004 can have the same function type.</p>
1005
1006<p>Parameter attributes are simple keywords that follow the type specified. If
1007 multiple parameter attributes are needed, they are space separated. For
1008 example:</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1009
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1010<pre class="doc_code">
Nick Lewyckyb6a7d252009-02-15 23:06:14 +0000[diff] [blame]1011declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner66d922c2008-10-04 18:33:34 +0000[diff] [blame]1012declare i32 @atoi(i8 zeroext)
1013declare signext i8 @returns_signed_char()
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1014</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1015
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1016<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1017 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1018
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1019<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1020
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1021<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1022 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1023 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichebe81732011-03-16 22:20:18 +0000[diff] [blame]1024 should be zero-extended to the extent required by the target's ABI (which
1025 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1026 parameter) or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1027
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1028 <dt><tt><b>signext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1029 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich9e69ff92011-03-17 14:21:58 +0000[diff] [blame]1030 should be sign-extended to the extent required by the target's ABI (which
1031 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1032 return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1033
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1034 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1035 <dd>This indicates that this parameter or return value should be treated in a
1036 special target-dependent fashion during while emitting code for a function
1037 call or return (usually, by putting it in a register as opposed to memory,
1038 though some targets use it to distinguish between two different kinds of
1039 registers). Use of this attribute is target-specific.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1040
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1041 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1042 <dd><p>This indicates that the pointer parameter should really be passed by
1043 value to the function. The attribute implies that a hidden copy of the
1044 pointee
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1045 is made between the caller and the callee, so the callee is unable to
1046 modify the value in the callee. This attribute is only valid on LLVM
1047 pointer arguments. It is generally used to pass structs and arrays by
1048 value, but is also valid on pointers to scalars. The copy is considered
1049 to belong to the caller not the callee (for example,
1050 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1051 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1052 values.</p>
1053
1054 <p>The byval attribute also supports specifying an alignment with
1055 the align attribute. It indicates the alignment of the stack slot to
1056 form and the known alignment of the pointer specified to the call site. If
1057 the alignment is not specified, then the code generator makes a
1058 target-specific assumption.</p></dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1059
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1060 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1061 <dd>This indicates that the pointer parameter specifies the address of a
1062 structure that is the return value of the function in the source program.
1063 This pointer must be guaranteed by the caller to be valid: loads and
1064 stores to the structure may be assumed by the callee to not to trap. This
1065 may only be applied to the first parameter. This is not a valid attribute
1066 for return values. </dd>
1067
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1068 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1069 <dd>This indicates that pointer values
1070 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmanefca7f92010-07-02 23:46:54 +0000[diff] [blame]1071 value do not alias pointer values which are not <i>based</i> on it,
1072 ignoring certain "irrelevant" dependencies.
1073 For a call to the parent function, dependencies between memory
1074 references from before or after the call and from those during the call
1075 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1076 return value used in that call.
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1077 The caller shares the responsibility with the callee for ensuring that
1078 these requirements are met.
1079 For further details, please see the discussion of the NoAlias response in
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1080 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1081<br>
John McCall191d4ee2010-07-06 21:07:14 +0000[diff] [blame]1082 Note that this definition of <tt>noalias</tt> is intentionally
1083 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner211244a2010-07-06 20:51:35 +0000[diff] [blame]1084 arguments, though it is slightly weaker.
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1085<br>
1086 For function return values, C99's <tt>restrict</tt> is not meaningful,
1087 while LLVM's <tt>noalias</tt> is.
1088 </dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1089
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1090 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1091 <dd>This indicates that the callee does not make any copies of the pointer
1092 that outlive the callee itself. This is not a valid attribute for return
1093 values.</dd>
1094
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1095 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1096 <dd>This indicates that the pointer parameter can be excised using the
1097 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1098 attribute for return values.</dd>
1099</dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1100
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1101</div>
1102
1103<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1104<h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1105 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1106</h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1107
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1108<div>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1109
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1110<p>Each function may specify a garbage collector name, which is simply a
1111 string:</p>
1112
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1113<pre class="doc_code">
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1114define void @f() gc "name" { ... }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1115</pre>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1116
1117<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1118 collector which will cause the compiler to alter its output in order to
1119 support the named garbage collection algorithm.</p>
1120
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1121</div>
1122
1123<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1124<h3>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1125 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1126</h3>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1127
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1128<div>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1129
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1130<p>Function attributes are set to communicate additional information about a
1131 function. Function attributes are considered to be part of the function, not
1132 of the function type, so functions with different parameter attributes can
1133 have the same function type.</p>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1134
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1135<p>Function attributes are simple keywords that follow the type specified. If
1136 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1137
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1138<pre class="doc_code">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1139define void @f() noinline { ... }
1140define void @f() alwaysinline { ... }
1141define void @f() alwaysinline optsize { ... }
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1142define void @f() optsize { ... }
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1143</pre>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1144
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1145<dl>
Kostya Serebryany164b86b2012-01-20 17:56:17 +0000[diff] [blame]1146 <dt><tt><b>address_safety</b></tt></dt>
1147 <dd>This attribute indicates that the address safety analysis
1148 is enabled for this function. </dd>
1149
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
Dan Gohman129bd562011-06-16 16:03:13 +0000[diff] [blame]1160 <dt><tt><b>nonlazybind</b></tt></dt>
1161 <dd>This attribute suppresses lazy symbol binding for the function. This
1162 may make calls to the function faster, at the cost of extra program
1163 startup time if the function is not called during program startup.</dd>
1164
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +0000[diff] [blame]1165 <dt><tt><b>inlinehint</b></tt></dt>
1166 <dd>This attribute indicates that the source code contained a hint that inlining
1167 this function is desirable (such as the "inline" keyword in C/C++). It
1168 is just a hint; it imposes no requirements on the inliner.</dd>
1169
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1170 <dt><tt><b>naked</b></tt></dt>
1171 <dd>This attribute disables prologue / epilogue emission for the function.
1172 This can have very system-specific consequences.</dd>
1173
1174 <dt><tt><b>noimplicitfloat</b></tt></dt>
1175 <dd>This attributes disables implicit floating point instructions.</dd>
1176
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1177 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1178 <dd>This attribute indicates that the inliner should never inline this
1179 function in any situation. This attribute may not be used together with
1180 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1181
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1182 <dt><tt><b>noredzone</b></tt></dt>
1183 <dd>This attribute indicates that the code generator should not use a red
1184 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1185
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1186 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1187 <dd>This function attribute indicates that the function never returns
1188 normally. This produces undefined behavior at runtime if the function
1189 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +0000[diff] [blame]1190
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1191 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1192 <dd>This function attribute indicates that the function never returns with an
1193 unwind or exceptional control flow. If the function does unwind, its
1194 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +0000[diff] [blame]1195
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1196 <dt><tt><b>optsize</b></tt></dt>
1197 <dd>This attribute suggests that optimization passes and code generator passes
1198 make choices that keep the code size of this function low, and otherwise
1199 do optimizations specifically to reduce code size.</dd>
1200
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1201 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1202 <dd>This attribute indicates that the function computes its result (or decides
1203 to unwind an exception) based strictly on its arguments, without
1204 dereferencing any pointer arguments or otherwise accessing any mutable
1205 state (e.g. memory, control registers, etc) visible to caller functions.
1206 It does not write through any pointer arguments
1207 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1208 changes any state visible to callers. This means that it cannot unwind
1209 exceptions by calling the <tt>C++</tt> exception throwing methods, but
1210 could use the <tt>unwind</tt> instruction.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +0000[diff] [blame]1211
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1212 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1213 <dd>This attribute indicates that the function does not write through any
1214 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1215 arguments) or otherwise modify any state (e.g. memory, control registers,
1216 etc) visible to caller functions. It may dereference pointer arguments
1217 and read state that may be set in the caller. A readonly function always
1218 returns the same value (or unwinds an exception identically) when called
1219 with the same set of arguments and global state. It cannot unwind an
1220 exception by calling the <tt>C++</tt> exception throwing methods, but may
1221 use the <tt>unwind</tt> instruction.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +0000[diff] [blame]1222
Bill Wendling9bd5d042011-12-05 21:27:54 +0000[diff] [blame]1223 <dt><tt><b><a name="returns_twice">returns_twice</a></b></tt></dt>
1224 <dd>This attribute indicates that this function can return twice. The
1225 C <code>setjmp</code> is an example of such a function. The compiler
1226 disables some optimizations (like tail calls) in the caller of these
1227 functions.</dd>
1228
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1229 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1230 <dd>This attribute indicates that the function should emit a stack smashing
1231 protector. It is in the form of a "canary"&mdash;a random value placed on
1232 the stack before the local variables that's checked upon return from the
1233 function to see if it has been overwritten. A heuristic is used to
1234 determine if a function needs stack protectors or not.<br>
1235<br>
1236 If a function that has an <tt>ssp</tt> attribute is inlined into a
1237 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1238 function will have an <tt>ssp</tt> attribute.</dd>
1239
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1240 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1241 <dd>This attribute indicates that the function should <em>always</em> emit a
1242 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1243 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1244<br>
1245 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1246 function that doesn't have an <tt>sspreq</tt> attribute or which has
1247 an <tt>ssp</tt> attribute, then the resulting function will have
1248 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindolafbff0ec2011-07-25 15:27:59 +0000[diff] [blame]1249
1250 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1251 <dd>This attribute indicates that the ABI being targeted requires that
1252 an unwind table entry be produce for this function even if we can
1253 show that no exceptions passes by it. This is normally the case for
1254 the ELF x86-64 abi, but it can be disabled for some compilation
1255 units.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1256</dl>
1257
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1258</div>
1259
1260<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1261<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]1262 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1263</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1264
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1265<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1266
1267<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1268 the GCC "file scope inline asm" blocks. These blocks are internally
1269 concatenated by LLVM and treated as a single unit, but may be separated in
1270 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1271
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1272<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1273module asm "inline asm code goes here"
1274module asm "more can go here"
1275</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1276
1277<p>The strings can contain any character by escaping non-printable characters.
1278 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]1279 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1280
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1281<p>The inline asm code is simply printed to the machine code .s file when
1282 assembly code is generated.</p>
1283
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1284</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1285
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1286<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1287<h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1288 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1289</h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1290
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1291<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1292
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1293<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1294 data is to be laid out in memory. The syntax for the data layout is
1295 simply:</p>
1296
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1297<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1298target datalayout = "<i>layout specification</i>"
1299</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1300
1301<p>The <i>layout specification</i> consists of a list of specifications
1302 separated by the minus sign character ('-'). Each specification starts with
1303 a letter and may include other information after the letter to define some
1304 aspect of the data layout. The specifications accepted are as follows:</p>
1305
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1306<dl>
1307 <dt><tt>E</tt></dt>
1308 <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]1309 bits with the most significance have the lowest address location.</dd>
1310
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1311 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1312 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1313 the bits with the least significance have the lowest address
1314 location.</dd>
1315
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1316 <dt><tt>S<i>size</i></tt></dt>
1317 <dd>Specifies the natural alignment of the stack in bits. Alignment promotion
1318 of stack variables is limited to the natural stack alignment to avoid
1319 dynamic stack realignment. The stack alignment must be a multiple of
Lang Hames5f119a62011-10-11 17:50:14 +0000[diff] [blame]1320 8-bits. If omitted, the natural stack alignment defaults to "unspecified",
1321 which does not prevent any alignment promotions.</dd>
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1322
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1323 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1324 <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]1325 <i>preferred</i> alignments. All sizes are in bits. Specifying
1326 the <i>pref</i> alignment is optional. If omitted, the
1327 preceding <tt>:</tt> should be omitted too.</dd>
1328
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1329 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1330 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1331 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1332
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1333 <dt><tt>v<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 vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1335 <i>size</i>.</dd>
1336
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1337 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1338 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +0000[diff] [blame]1339 <i>size</i>. Only values of <i>size</i> that are supported by the target
1340 will work. 32 (float) and 64 (double) are supported on all targets;
1341 80 or 128 (different flavors of long double) are also supported on some
1342 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1343
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1344 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1345 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1346 <i>size</i>.</dd>
1347
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1348 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1349 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1350 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1351
1352 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1353 <dd>This specifies a set of native integer widths for the target CPU
1354 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1355 "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]1356 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1357 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1358</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1359
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1360<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]1361 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1362 specifications in the <tt>datalayout</tt> keyword. The default specifications
1363 are given in this list:</p>
1364
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1365<ul>
1366 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +0000[diff] [blame]1367 <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]1368 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1369 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1370 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1371 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1372 <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]1373 alignment of 64-bits</li>
1374 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1375 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1376 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1377 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1378 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1379 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1380</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1381
1382<p>When LLVM is determining the alignment for a given type, it uses the
1383 following rules:</p>
1384
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1385<ol>
1386 <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]1387 specification is used.</li>
1388
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1389 <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]1390 smallest integer type that is larger than the bitwidth of the sought type
1391 is used. If none of the specifications are larger than the bitwidth then
1392 the the largest integer type is used. For example, given the default
1393 specifications above, the i7 type will use the alignment of i8 (next
1394 largest) while both i65 and i256 will use the alignment of i64 (largest
1395 specified).</li>
1396
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1397 <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]1398 largest vector type that is smaller than the sought vector type will be
1399 used as a fall back. This happens because &lt;128 x double&gt; can be
1400 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1401</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1402
Chris Lattner6509f502011-10-11 23:01:39 +0000[diff] [blame]1403<p>The function of the data layout string may not be what you expect. Notably,
1404 this is not a specification from the frontend of what alignment the code
1405 generator should use.</p>
1406
1407<p>Instead, if specified, the target data layout is required to match what the
1408 ultimate <em>code generator</em> expects. This string is used by the
1409 mid-level optimizers to
1410 improve code, and this only works if it matches what the ultimate code
1411 generator uses. If you would like to generate IR that does not embed this
1412 target-specific detail into the IR, then you don't have to specify the
1413 string. This will disable some optimizations that require precise layout
1414 information, but this also prevents those optimizations from introducing
1415 target specificity into the IR.</p>
1416
1417
1418
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1419</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1420
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1421<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1422<h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1423 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1424</h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1425
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1426<div>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1427
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1428<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +0000[diff] [blame]1429with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1430is undefined. Pointer values are associated with address ranges
1431according to the following rules:</p>
1432
1433<ul>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1434 <li>A pointer value is associated with the addresses associated with
1435 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1436 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1437 range of the variable's storage.</li>
1438 <li>The result value of an allocation instruction is associated with
1439 the address range of the allocated storage.</li>
1440 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1441 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1442 <li>An integer constant other than zero or a pointer value returned
1443 from a function not defined within LLVM may be associated with address
1444 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1445 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1446 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1447</ul>
1448
1449<p>A pointer value is <i>based</i> on another pointer value according
1450 to the following rules:</p>
1451
1452<ul>
1453 <li>A pointer value formed from a
1454 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1455 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1456 <li>The result value of a
1457 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1458 of the <tt>bitcast</tt>.</li>
1459 <li>A pointer value formed by an
1460 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1461 pointer values that contribute (directly or indirectly) to the
1462 computation of the pointer's value.</li>
1463 <li>The "<i>based</i> on" relationship is transitive.</li>
1464</ul>
1465
1466<p>Note that this definition of <i>"based"</i> is intentionally
1467 similar to the definition of <i>"based"</i> in C99, though it is
1468 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1469
1470<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1471<tt><a href="#i_load">load</a></tt> merely indicates the size and
1472alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +0000[diff] [blame]1473interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1474<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1475and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1476
1477<p>Consequently, type-based alias analysis, aka TBAA, aka
1478<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1479LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1480additional information which specialized optimization passes may use
1481to implement type-based alias analysis.</p>
1482
1483</div>
1484
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1485<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1486<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1487 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1488</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1489
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1490<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1491
1492<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1493href="#i_store"><tt>store</tt></a>s, and <a
1494href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1495The optimizers must not change the number of volatile operations or change their
1496order of execution relative to other volatile operations. The optimizers
1497<i>may</i> change the order of volatile operations relative to non-volatile
1498operations. This is not Java's "volatile" and has no cross-thread
1499synchronization behavior.</p>
1500
1501</div>
1502
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1503<!-- ======================================================================= -->
1504<h3>
1505 <a name="memmodel">Memory Model for Concurrent Operations</a>
1506</h3>
1507
1508<div>
1509
1510<p>The LLVM IR does not define any way to start parallel threads of execution
1511or to register signal handlers. Nonetheless, there are platform-specific
1512ways to create them, and we define LLVM IR's behavior in their presence. This
1513model is inspired by the C++0x memory model.</p>
1514
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1515<p>For a more informal introduction to this model, see the
1516<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.
1517
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1518<p>We define a <i>happens-before</i> partial order as the least partial order
1519that</p>
1520<ul>
1521 <li>Is a superset of single-thread program order, and</li>
1522 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1523 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1524 by platform-specific techniques, like pthread locks, thread
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1525 creation, thread joining, etc., and by atomic instructions.
1526 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1527 </li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1528</ul>
1529
1530<p>Note that program order does not introduce <i>happens-before</i> edges
1531between a thread and signals executing inside that thread.</p>
1532
1533<p>Every (defined) read operation (load instructions, memcpy, atomic
1534loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1535(defined) write operations (store instructions, atomic
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1536stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1537initialized globals are considered to have a write of the initializer which is
1538atomic and happens before any other read or write of the memory in question.
1539For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1540any write to the same byte, except:</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1541
1542<ul>
1543 <li>If <var>write<sub>1</sub></var> happens before
1544 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1545 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1546 does not see <var>write<sub>1</sub></var>.
Bill Wendling0246bb72011-07-31 06:45:03 +0000[diff] [blame]1547 <li>If <var>R<sub>byte</sub></var> happens before
1548 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1549 see <var>write<sub>3</sub></var>.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1550</ul>
1551
1552<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1553<ul>
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1554 <li>If <var>R</var> is volatile, the result is target-dependent. (Volatile
1555 is supposed to give guarantees which can support
1556 <code>sig_atomic_t</code> in C/C++, and may be used for accesses to
1557 addresses which do not behave like normal memory. It does not generally
1558 provide cross-thread synchronization.)
1559 <li>Otherwise, if there is no write to the same byte that happens before
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1560 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1561 <tt>undef</tt> for that byte.
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1562 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1563 <var>R<sub>byte</sub></var> returns the value written by that
1564 write.</li>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1565 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1566 <var>R<sub>byte</sub></var> may see are atomic, it chooses one of the
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1567 values written. See the <a href="#ordering">Atomic Memory Ordering
1568 Constraints</a> section for additional constraints on how the choice
1569 is made.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1570 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1571</ul>
1572
1573<p><var>R</var> returns the value composed of the series of bytes it read.
1574This implies that some bytes within the value may be <tt>undef</tt>
1575<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1576defines the semantics of the operation; it doesn't mean that targets will
1577emit more than one instruction to read the series of bytes.</p>
1578
1579<p>Note that in cases where none of the atomic intrinsics are used, this model
1580places only one restriction on IR transformations on top of what is required
1581for single-threaded execution: introducing a store to a byte which might not
Eli Friedman101c81d2011-08-02 01:15:34 +0000[diff] [blame]1582otherwise be stored is not allowed in general. (Specifically, in the case
1583where another thread might write to and read from an address, introducing a
1584store can change a load that may see exactly one write into a load that may
1585see multiple writes.)</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1586
1587<!-- FIXME: This model assumes all targets where concurrency is relevant have
1588a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1589none of the backends currently in the tree fall into this category; however,
1590there might be targets which care. If there are, we want a paragraph
1591like the following:
1592
1593Targets may specify that stores narrower than a certain width are not
1594available; on such a target, for the purposes of this model, treat any
1595non-atomic write with an alignment or width less than the minimum width
1596as if it writes to the relevant surrounding bytes.
1597-->
1598
1599</div>
1600
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1601<!-- ======================================================================= -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1602<h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1603 <a name="ordering">Atomic Memory Ordering Constraints</a>
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1604</h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1605
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1606<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1607
1608<p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]1609<a href="#i_atomicrmw"><code>atomicrmw</code></a>,
1610<a href="#i_fence"><code>fence</code></a>,
1611<a href="#i_load"><code>atomic load</code></a>, and
Eli Friedman8fa281a2011-08-09 23:26:12 +0000[diff] [blame]1612<a href="#i_store"><code>atomic store</code></a>) take an ordering parameter
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1613that determines which other atomic instructions on the same address they
1614<i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1615but are somewhat more colloquial. If these descriptions aren't precise enough,
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1616check those specs (see spec references in the
Nick Lewycky300a2632012-01-23 08:47:21 +0000[diff] [blame]1617<a href="Atomics.html#introduction">atomics guide</a>).
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1618<a href="#i_fence"><code>fence</code></a> instructions
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1619treat these orderings somewhat differently since they don't take an address.
1620See that instruction's documentation for details.</p>
1621
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1622<p>For a simpler introduction to the ordering constraints, see the
1623<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.</p>
1624
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1625<dl>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1626<dt><code>unordered</code></dt>
1627<dd>The set of values that can be read is governed by the happens-before
1628partial order. A value cannot be read unless some operation wrote it.
1629This is intended to provide a guarantee strong enough to model Java's
1630non-volatile shared variables. This ordering cannot be specified for
1631read-modify-write operations; it is not strong enough to make them atomic
1632in any interesting way.</dd>
1633<dt><code>monotonic</code></dt>
1634<dd>In addition to the guarantees of <code>unordered</code>, there is a single
1635total order for modifications by <code>monotonic</code> operations on each
1636address. All modification orders must be compatible with the happens-before
1637order. There is no guarantee that the modification orders can be combined to
1638a global total order for the whole program (and this often will not be
1639possible). The read in an atomic read-modify-write operation
1640(<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1641<a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1642reads the value in the modification order immediately before the value it
1643writes. If one atomic read happens before another atomic read of the same
1644address, the later read must see the same value or a later value in the
1645address's modification order. This disallows reordering of
1646<code>monotonic</code> (or stronger) operations on the same address. If an
1647address is written <code>monotonic</code>ally by one thread, and other threads
1648<code>monotonic</code>ally read that address repeatedly, the other threads must
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1649eventually see the write. This corresponds to the C++0x/C1x
1650<code>memory_order_relaxed</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1651<dt><code>acquire</code></dt>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1652<dd>In addition to the guarantees of <code>monotonic</code>,
Eli Friedmanc264b2f2011-08-24 20:28:39 +0000[diff] [blame]1653a <i>synchronizes-with</i> edge may be formed with a <code>release</code>
1654operation. This is intended to model C++'s <code>memory_order_acquire</code>.</dd>
1655<dt><code>release</code></dt>
1656<dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1657writes a value which is subsequently read by an <code>acquire</code> operation,
1658it <i>synchronizes-with</i> that operation. (This isn't a complete
1659description; see the C++0x definition of a release sequence.) This corresponds
1660to the C++0x/C1x <code>memory_order_release</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1661<dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1662<code>acquire</code> and <code>release</code> operation on its address.
1663This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1664<dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1665<dd>In addition to the guarantees of <code>acq_rel</code>
1666(<code>acquire</code> for an operation which only reads, <code>release</code>
1667for an operation which only writes), there is a global total order on all
1668sequentially-consistent operations on all addresses, which is consistent with
1669the <i>happens-before</i> partial order and with the modification orders of
1670all the affected addresses. Each sequentially-consistent read sees the last
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1671preceding write to the same address in this global order. This corresponds
1672to the C++0x/C1x <code>memory_order_seq_cst</code> and Java volatile.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1673</dl>
1674
1675<p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1676it only <i>synchronizes with</i> or participates in modification and seq_cst
1677total orderings with other operations running in the same thread (for example,
1678in signal handlers).</p>
1679
1680</div>
1681
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1682</div>
1683
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1684<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1685<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1686<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1687
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1688<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1689
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1690<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1691 intermediate representation. Being typed enables a number of optimizations
1692 to be performed on the intermediate representation directly, without having
1693 to do extra analyses on the side before the transformation. A strong type
1694 system makes it easier to read the generated code and enables novel analyses
1695 and transformations that are not feasible to perform on normal three address
1696 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1697
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1698<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1699<h3>
1700 <a name="t_classifications">Type Classifications</a>
1701</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1702
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1703<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1704
1705<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1706
1707<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1708 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1709 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1710 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1711 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1712 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1713 </tr>
1714 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1715 <td><a href="#t_floating">floating point</a></td>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1716 <td><tt>half, float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1717 </tr>
1718 <tr>
1719 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1720 <td><a href="#t_integer">integer</a>,
1721 <a href="#t_floating">floating point</a>,
1722 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +0000[diff] [blame]1723 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]1724 <a href="#t_struct">structure</a>,
1725 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1726 <a href="#t_label">label</a>,
1727 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1728 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1729 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1730 <tr>
1731 <td><a href="#t_primitive">primitive</a></td>
1732 <td><a href="#t_label">label</a>,
1733 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +0000[diff] [blame]1734 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1735 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1736 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1737 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1738 </tr>
1739 <tr>
1740 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1741 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1742 <a href="#t_function">function</a>,
1743 <a href="#t_pointer">pointer</a>,
1744 <a href="#t_struct">structure</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1745 <a href="#t_vector">vector</a>,
1746 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +0000[diff] [blame]1747 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1748 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1749 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1750</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1751
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1752<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1753 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1754 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1755
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1756</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1757
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1758<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1759<h3>
1760 <a name="t_primitive">Primitive Types</a>
1761</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1762
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1763<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1764
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1765<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1766 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1767
1768<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1769<h4>
1770 <a name="t_integer">Integer Type</a>
1771</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1772
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1773<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1774
1775<h5>Overview:</h5>
1776<p>The integer type is a very simple type that simply specifies an arbitrary
1777 bit width for the integer type desired. Any bit width from 1 bit to
1778 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1779
1780<h5>Syntax:</h5>
1781<pre>
1782 iN
1783</pre>
1784
1785<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1786 value.</p>
1787
1788<h5>Examples:</h5>
1789<table class="layout">
1790 <tr class="layout">
1791 <td class="left"><tt>i1</tt></td>
1792 <td class="left">a single-bit integer.</td>
1793 </tr>
1794 <tr class="layout">
1795 <td class="left"><tt>i32</tt></td>
1796 <td class="left">a 32-bit integer.</td>
1797 </tr>
1798 <tr class="layout">
1799 <td class="left"><tt>i1942652</tt></td>
1800 <td class="left">a really big integer of over 1 million bits.</td>
1801 </tr>
1802</table>
1803
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1804</div>
1805
1806<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1807<h4>
1808 <a name="t_floating">Floating Point Types</a>
1809</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1810
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1811<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1812
1813<table>
1814 <tbody>
1815 <tr><th>Type</th><th>Description</th></tr>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1816 <tr><td><tt>half</tt></td><td>16-bit floating point value</td></tr>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1817 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1818 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1819 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1820 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1821 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1822 </tbody>
1823</table>
1824
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1825</div>
1826
1827<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1828<h4>
1829 <a name="t_x86mmx">X86mmx Type</a>
1830</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1831
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1832<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1833
1834<h5>Overview:</h5>
1835<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>
1836
1837<h5>Syntax:</h5>
1838<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +0000[diff] [blame]1839 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1840</pre>
1841
1842</div>
1843
1844<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1845<h4>
1846 <a name="t_void">Void Type</a>
1847</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1848
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1849<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1850
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1851<h5>Overview:</h5>
1852<p>The void type does not represent any value and has no size.</p>
1853
1854<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1855<pre>
1856 void
1857</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1858
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1859</div>
1860
1861<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1862<h4>
1863 <a name="t_label">Label Type</a>
1864</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1865
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1866<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1867
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1868<h5>Overview:</h5>
1869<p>The label type represents code labels.</p>
1870
1871<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1872<pre>
1873 label
1874</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1875
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1876</div>
1877
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1878<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1879<h4>
1880 <a name="t_metadata">Metadata Type</a>
1881</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1882
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1883<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1884
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1885<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1886<p>The metadata type represents embedded metadata. No derived types may be
1887 created from metadata except for <a href="#t_function">function</a>
1888 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1889
1890<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1891<pre>
1892 metadata
1893</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1894
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1895</div>
1896
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1897</div>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1898
1899<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1900<h3>
1901 <a name="t_derived">Derived Types</a>
1902</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1903
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1904<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1905
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1906<p>The real power in LLVM comes from the derived types in the system. This is
1907 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1908 useful types. Each of these types contain one or more element types which
1909 may be a primitive type, or another derived type. For example, it is
1910 possible to have a two dimensional array, using an array as the element type
1911 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1912
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1913<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1914<h4>
1915 <a name="t_aggregate">Aggregate Types</a>
1916</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1917
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1918<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1919
1920<p>Aggregate Types are a subset of derived types that can contain multiple
Duncan Sands20536b52011-12-14 15:44:20 +0000[diff] [blame]1921 member types. <a href="#t_array">Arrays</a> and
1922 <a href="#t_struct">structs</a> are aggregate types.
1923 <a href="#t_vector">Vectors</a> are not considered to be aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1924
1925</div>
1926
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1927<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1928<h4>
1929 <a name="t_array">Array Type</a>
1930</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1931
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1932<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1933
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1934<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1935<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1936 sequentially in memory. The array type requires a size (number of elements)
1937 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1938
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1939<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1940<pre>
1941 [&lt;# elements&gt; x &lt;elementtype&gt;]
1942</pre>
1943
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1944<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1945 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1946
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1947<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1948<table class="layout">
1949 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1950 <td class="left"><tt>[40 x i32]</tt></td>
1951 <td class="left">Array of 40 32-bit integer values.</td>
1952 </tr>
1953 <tr class="layout">
1954 <td class="left"><tt>[41 x i32]</tt></td>
1955 <td class="left">Array of 41 32-bit integer values.</td>
1956 </tr>
1957 <tr class="layout">
1958 <td class="left"><tt>[4 x i8]</tt></td>
1959 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1960 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1961</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1962<p>Here are some examples of multidimensional arrays:</p>
1963<table class="layout">
1964 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1965 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1966 <td class="left">3x4 array of 32-bit integer values.</td>
1967 </tr>
1968 <tr class="layout">
1969 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1970 <td class="left">12x10 array of single precision floating point values.</td>
1971 </tr>
1972 <tr class="layout">
1973 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1974 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1975 </tr>
1976</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1977
Dan Gohman7657f6b2009-11-09 19:01:53 +0000[diff] [blame]1978<p>There is no restriction on indexing beyond the end of the array implied by
1979 a static type (though there are restrictions on indexing beyond the bounds
1980 of an allocated object in some cases). This means that single-dimension
1981 'variable sized array' addressing can be implemented in LLVM with a zero
1982 length array type. An implementation of 'pascal style arrays' in LLVM could
1983 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1984
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1985</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1986
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1987<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1988<h4>
1989 <a name="t_function">Function Type</a>
1990</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1991
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1992<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]1993
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1994<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1995<p>The function type can be thought of as a function signature. It consists of
1996 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]1997 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]1998
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1999<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2000<pre>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2001 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2002</pre>
2003
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2004<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]2005 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
2006 which indicates that the function takes a variable number of arguments.
2007 Variable argument functions can access their arguments with
2008 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2009 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]2010 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2011
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2012<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2013<table class="layout">
2014 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2015 <td class="left"><tt>i32 (i32)</tt></td>
2016 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2017 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2018 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2019 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]2020 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2021 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2022 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
2023 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2024 </td>
2025 </tr><tr class="layout">
2026 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2027 <td class="left">A vararg function that takes at least one
2028 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2029 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2030 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2031 </td>
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2032 </tr><tr class="layout">
2033 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2034 <td class="left">A function taking an <tt>i32</tt>, returning a
2035 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2036 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2037 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2038</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2039
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2040</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2041
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2042<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2043<h4>
2044 <a name="t_struct">Structure Type</a>
2045</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2046
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2047<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2048
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2049<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2050<p>The structure type is used to represent a collection of data members together
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2051 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]2052
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]2053<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2054 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2055 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2056 Structures in registers are accessed using the
2057 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2058 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2059
2060<p>Structures may optionally be "packed" structures, which indicate that the
2061 alignment of the struct is one byte, and that there is no padding between
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2062 the elements. In non-packed structs, padding between field types is inserted
2063 as defined by the TargetData string in the module, which is required to match
Chris Lattnere4617b02011-10-11 23:02:17 +0000[diff] [blame]2064 what the underlying code generator expects.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2065
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2066<p>Structures can either be "literal" or "identified". A literal structure is
2067 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) whereas identified
2068 types are always defined at the top level with a name. Literal types are
2069 uniqued by their contents and can never be recursive or opaque since there is
Chris Lattneraa175c32011-08-12 18:12:40 +0000[diff] [blame]2070 no way to write one. Identified types can be recursive, can be opaqued, and are
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2071 never uniqued.
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2072</p>
2073
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2074<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2075<pre>
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2076 %T1 = type { &lt;type list&gt; } <i>; Identified normal struct type</i>
2077 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Identified packed struct type</i>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2078</pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2079
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2080<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2081<table class="layout">
2082 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2083 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2084 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2085 </tr>
2086 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2087 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
2088 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2089 second element is a <a href="#t_pointer">pointer</a> to a
2090 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2091 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2092 </tr>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2093 <tr class="layout">
2094 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
2095 <td class="left">A packed struct known to be 5 bytes in size.</td>
2096 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2097</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2098
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2099</div>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2100
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2101<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2102<h4>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2103 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2104</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2105
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2106<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2107
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2108<h5>Overview:</h5>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2109<p>Opaque structure types are used to represent named structure types that do
2110 not have a body specified. This corresponds (for example) to the C notion of
2111 a forward declared structure.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2112
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2113<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2114<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2115 %X = type opaque
2116 %52 = type opaque
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2117</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2118
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2119<h5>Examples:</h5>
2120<table class="layout">
2121 <tr class="layout">
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2122 <td class="left"><tt>opaque</tt></td>
2123 <td class="left">An opaque type.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2124 </tr>
2125</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2126
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2127</div>
2128
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2129
2130
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2131<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2132<h4>
2133 <a name="t_pointer">Pointer Type</a>
2134</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2135
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2136<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2137
2138<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +0000[diff] [blame]2139<p>The pointer type is used to specify memory locations.
2140 Pointers are commonly used to reference objects in memory.</p>
2141
2142<p>Pointer types may have an optional address space attribute defining the
2143 numbered address space where the pointed-to object resides. The default
2144 address space is number zero. The semantics of non-zero address
2145 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2146
2147<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2148 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2149
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2150<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2151<pre>
2152 &lt;type&gt; *
2153</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2154
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2155<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2156<table class="layout">
2157 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2158 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2159 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2160 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2161 </tr>
2162 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]2163 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2164 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2165 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2166 <tt>i32</tt>.</td>
2167 </tr>
2168 <tr class="layout">
2169 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2170 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2171 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2172 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2173</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2174
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2175</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2176
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2177<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2178<h4>
2179 <a name="t_vector">Vector Type</a>
2180</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2181
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2182<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2183
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2184<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2185<p>A vector type is a simple derived type that represents a vector of elements.
2186 Vector types are used when multiple primitive data are operated in parallel
2187 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +0000[diff] [blame]2188 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2189 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2190
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2191<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2192<pre>
2193 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2194</pre>
2195
Chris Lattner7d2e7be2010-10-10 18:20:35 +0000[diff] [blame]2196<p>The number of elements is a constant integer value larger than 0; elementtype
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2197 may be any integer or floating point type, or a pointer to these types.
2198 Vectors of size zero are not allowed. </p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2199
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2200<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2201<table class="layout">
2202 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2203 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2204 <td class="left">Vector of 4 32-bit integer values.</td>
2205 </tr>
2206 <tr class="layout">
2207 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2208 <td class="left">Vector of 8 32-bit floating-point values.</td>
2209 </tr>
2210 <tr class="layout">
2211 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2212 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2213 </tr>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2214 <tr class="layout">
2215 <td class="left"><tt>&lt;4 x i64*&gt;</tt></td>
2216 <td class="left">Vector of 4 pointers to 64-bit integer values.</td>
2217 </tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2218</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2219
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2220</div>
2221
Bill Wendlingaf75f0c2011-07-31 06:47:33 +0000[diff] [blame]2222</div>
2223
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]2224</div>
2225
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2226<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2227<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2228<!-- *********************************************************************** -->
2229
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2230<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2231
2232<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2233 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2234
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2235<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2236<h3>
2237 <a name="simpleconstants">Simple Constants</a>
2238</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2239
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2240<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2241
2242<dl>
2243 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2244 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]2245 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2246
2247 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2248 <dd>Standard integers (such as '4') are constants of
2249 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2250 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2251
2252 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2253 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2254 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2255 notation (see below). The assembler requires the exact decimal value of a
2256 floating-point constant. For example, the assembler accepts 1.25 but
2257 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2258 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2259
2260 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2261 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2262 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2263</dl>
2264
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2265<p>The one non-intuitive notation for constants is the hexadecimal form of
2266 floating point constants. For example, the form '<tt>double
2267 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2268 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2269 constants are required (and the only time that they are generated by the
2270 disassembler) is when a floating point constant must be emitted but it cannot
2271 be represented as a decimal floating point number in a reasonable number of
2272 digits. For example, NaN's, infinities, and other special values are
2273 represented in their IEEE hexadecimal format so that assembly and disassembly
2274 do not cause any bits to change in the constants.</p>
2275
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2276<p>When using the hexadecimal form, constants of types half, float, and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2277 represented using the 16-digit form shown above (which matches the IEEE754
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2278 representation for double); half and float values must, however, be exactly
2279 representable as IEE754 half and single precision, respectively.
2280 Hexadecimal format is always used
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2281 for long double, and there are three forms of long double. The 80-bit format
2282 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2283 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2284 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2285 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2286 currently supported target uses this format. Long doubles will only work if
2287 they match the long double format on your target. All hexadecimal formats
2288 are big-endian (sign bit at the left).</p>
2289
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]2290<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2291</div>
2292
2293<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2294<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]2295<a name="aggregateconstants"></a> <!-- old anchor -->
2296<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2297</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2298
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2299<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2300
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2301<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2302 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2303
2304<dl>
2305 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2306 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2307 type definitions (a comma separated list of elements, surrounded by braces
2308 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2309 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2310 Structure constants must have <a href="#t_struct">structure type</a>, and
2311 the number and types of elements must match those specified by the
2312 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2313
2314 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2315 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2316 definitions (a comma separated list of elements, surrounded by square
2317 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2318 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2319 the number and types of elements must match those specified by the
2320 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2321
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2322 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2323 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2324 definitions (a comma separated list of elements, surrounded by
2325 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2326 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2327 have <a href="#t_vector">vector type</a>, and the number and types of
2328 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2329
2330 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2331 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]2332 value to zero of <em>any</em> type, including scalar and
2333 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2334 This is often used to avoid having to print large zero initializers
2335 (e.g. for large arrays) and is always exactly equivalent to using explicit
2336 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +0000[diff] [blame]2337
2338 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +0000[diff] [blame]2339 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2340 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2341 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2342 be interpreted as part of the instruction stream, metadata is a place to
2343 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2344</dl>
2345
2346</div>
2347
2348<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2349<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2350 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2351</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2352
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2353<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2354
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2355<p>The addresses of <a href="#globalvars">global variables</a>
2356 and <a href="#functionstructure">functions</a> are always implicitly valid
2357 (link-time) constants. These constants are explicitly referenced when
2358 the <a href="#identifiers">identifier for the global</a> is used and always
2359 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2360 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2361
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2362<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]2363@X = global i32 17
2364@Y = global i32 42
2365@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2366</pre>
2367
2368</div>
2369
2370<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2371<h3>
2372 <a name="undefvalues">Undefined Values</a>
2373</h3>
2374
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2375<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2376
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2377<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]2378 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2379 Undefined values may be of any type (other than '<tt>label</tt>'
2380 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2381
Chris Lattnerc608cb12009-09-11 01:49:31 +0000[diff] [blame]2382<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2383 program is well defined no matter what value is used. This gives the
2384 compiler more freedom to optimize. Here are some examples of (potentially
2385 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2386
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2387
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2388<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2389 %A = add %X, undef
2390 %B = sub %X, undef
2391 %C = xor %X, undef
2392Safe:
2393 %A = undef
2394 %B = undef
2395 %C = undef
2396</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2397
2398<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]2399 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]2400
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2401<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2402 %A = or %X, undef
2403 %B = and %X, undef
2404Safe:
2405 %A = -1
2406 %B = 0
2407Unsafe:
2408 %A = undef
2409 %B = undef
2410</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2411
2412<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2413 For example, if <tt>%X</tt> has a zero bit, then the output of the
2414 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2415 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2416 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2417 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2418 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2419 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2420 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2421
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2422<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2423 %A = select undef, %X, %Y
2424 %B = select undef, 42, %Y
2425 %C = select %X, %Y, undef
2426Safe:
2427 %A = %X (or %Y)
2428 %B = 42 (or %Y)
2429 %C = %Y
2430Unsafe:
2431 %A = undef
2432 %B = undef
2433 %C = undef
2434</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2435
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2436<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2437 branch) conditions can go <em>either way</em>, but they have to come from one
2438 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2439 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2440 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2441 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2442 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2443 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2444
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2445<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2446 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2447
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2448 %B = undef
2449 %C = xor %B, %B
2450
2451 %D = undef
2452 %E = icmp lt %D, 4
2453 %F = icmp gte %D, 4
2454
2455Safe:
2456 %A = undef
2457 %B = undef
2458 %C = undef
2459 %D = undef
2460 %E = undef
2461 %F = undef
2462</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2463
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2464<p>This example points out that two '<tt>undef</tt>' operands are not
2465 necessarily the same. This can be surprising to people (and also matches C
2466 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2467 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2468 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2469 its value over its "live range". This is true because the variable doesn't
2470 actually <em>have a live range</em>. Instead, the value is logically read
2471 from arbitrary registers that happen to be around when needed, so the value
2472 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2473 need to have the same semantics or the core LLVM "replace all uses with"
2474 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2475
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2476<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2477 %A = fdiv undef, %X
2478 %B = fdiv %X, undef
2479Safe:
2480 %A = undef
2481b: unreachable
2482</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2483
2484<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2485 value</em> and <em>undefined behavior</em>. An undefined value (like
2486 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2487 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2488 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2489 defined on SNaN's. However, in the second example, we can make a more
2490 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2491 arbitrary value, we are allowed to assume that it could be zero. Since a
2492 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2493 the operation does not execute at all. This allows us to delete the divide and
2494 all code after it. Because the undefined operation "can't happen", the
2495 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2496
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2497<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2498a: store undef -> %X
2499b: store %X -> undef
2500Safe:
2501a: &lt;deleted&gt;
2502b: unreachable
2503</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2504
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2505<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2506 undefined value can be assumed to not have any effect; we can assume that the
2507 value is overwritten with bits that happen to match what was already there.
2508 However, a store <em>to</em> an undefined location could clobber arbitrary
2509 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2510
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2511</div>
2512
2513<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2514<h3>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2515 <a name="poisonvalues">Poison Values</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2516</h3>
2517
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2518<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2519
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2520<p>Poison values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2521 they also represent the fact that an instruction or constant expression which
2522 cannot evoke side effects has nevertheless detected a condition which results
2523 in undefined behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2524
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2525<p>There is currently no way of representing a poison value in the IR; they
Dan Gohman855abed2010-05-03 14:51:43 +0000[diff] [blame]2526 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2527 <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]2528
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2529<p>Poison value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2530
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2531<ul>
2532<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2533 their operands.</li>
2534
2535<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2536 to their dynamic predecessor basic block.</li>
2537
2538<li>Function arguments depend on the corresponding actual argument values in
2539 the dynamic callers of their functions.</li>
2540
2541<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2542 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2543 control back to them.</li>
2544
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2545<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
2546 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_unwind"><tt>unwind</tt></a>,
2547 or exception-throwing call instructions that dynamically transfer control
2548 back to them.</li>
2549
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2550<li>Non-volatile loads and stores depend on the most recent stores to all of the
2551 referenced memory addresses, following the order in the IR
2552 (including loads and stores implied by intrinsics such as
2553 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2554
Dan Gohman7c24ff12010-05-03 14:59:34 +0000[diff] [blame]2555<!-- TODO: In the case of multiple threads, this only applies if the store
2556 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2557
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2558<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2559
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2560<li>An instruction with externally visible side effects depends on the most
2561 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]2562 the order in the IR. (This includes
2563 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2564
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2565<li>An instruction <i>control-depends</i> on a
2566 <a href="#terminators">terminator instruction</a>
2567 if the terminator instruction has multiple successors and the instruction
2568 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2569 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2570
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2571<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2572 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2573 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2574 successor.</li>
2575
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2576<li>Dependence is transitive.</li>
2577
2578</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2579
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2580<p>Poison Values have the same behavior as <a href="#undefvalues">undef values</a>,
2581 with the additional affect that any instruction which has a <i>dependence</i>
2582 on a poison value has undefined behavior.</p>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2583
2584<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +0000[diff] [blame]2585
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2586<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2587entry:
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2588 %poison = sub nuw i32 0, 1 ; Results in a poison value.
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2589 %still_poison = and i32 %poison, 0 ; 0, but also poison.
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2590 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2591 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2592
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2593 store i32 %poison, i32* @g ; Poison value stored to memory.
2594 %poison2 = load i32* @g ; Poison value loaded back from memory.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2595
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2596 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2597
2598 %narrowaddr = bitcast i32* @g to i16*
2599 %wideaddr = bitcast i32* @g to i64*
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2600 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2601 %poison4 = load i64* %wideaddr ; Returns a poison value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2602
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2603 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2604 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2605
2606true:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2607 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2608 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2609 br label %end
2610
2611end:
2612 %p = phi i32 [ 0, %entry ], [ 1, %true ]
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2613 ; Both edges into this PHI are
2614 ; control-dependent on %cmp, so this
2615 ; always results in a poison value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2616
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2617 store volatile i32 0, i32* @g ; This would depend on the store in %true
2618 ; if %cmp is true, or the store in %entry
2619 ; otherwise, so this is undefined behavior.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2620
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2621 br i1 %cmp, label %second_true, label %second_end
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2622 ; The same branch again, but this time the
2623 ; true block doesn't have side effects.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2624
2625second_true:
2626 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2627 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2628
2629second_end:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2630 store volatile i32 0, i32* @g ; This time, the instruction always depends
2631 ; on the store in %end. Also, it is
2632 ; control-equivalent to %end, so this is
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2633 ; well-defined (ignoring earlier undefined
2634 ; behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2635</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2636
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2637</div>
2638
2639<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2640<h3>
2641 <a name="blockaddress">Addresses of Basic Blocks</a>
2642</h3>
2643
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2644<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2645
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2646<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2647
2648<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +0000[diff] [blame]2649 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2650 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2651
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2652<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2653 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2654 comparisons against null. Pointer equality tests between labels addresses
2655 results in undefined behavior &mdash; though, again, comparison against null
2656 is ok, and no label is equal to the null pointer. This may be passed around
2657 as an opaque pointer sized value as long as the bits are not inspected. This
2658 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2659 long as the original value is reconstituted before the <tt>indirectbr</tt>
2660 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2661
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2662<p>Finally, some targets may provide defined semantics when using the value as
2663 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2664
2665</div>
2666
2667
2668<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2669<h3>
2670 <a name="constantexprs">Constant Expressions</a>
2671</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2672
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2673<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2674
2675<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2676 to be used as constants. Constant expressions may be of
2677 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2678 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2679 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2680
2681<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2682 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2683 <dd>Truncate a constant to another type. The bit size of CST must be larger
2684 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2685
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2686 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2687 <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]2688 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2689
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2690 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2691 <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]2692 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2693
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2694 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2695 <dd>Truncate a floating point constant to another floating point type. The
2696 size of CST must be larger than the size of TYPE. Both types must be
2697 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2698
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2699 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2700 <dd>Floating point extend a constant to another type. The size of CST must be
2701 smaller or equal to the size of TYPE. Both types must be floating
2702 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2703
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2704 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2705 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2706 constant. TYPE must be a scalar or vector integer type. CST must be of
2707 scalar or vector floating point type. Both CST and TYPE must be scalars,
2708 or vectors of the same number of elements. If the value won't fit in the
2709 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2710
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2711 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2712 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2713 constant. TYPE must be a scalar or vector integer type. CST must be of
2714 scalar or vector floating point type. Both CST and TYPE must be scalars,
2715 or vectors of the same number of elements. If the value won't fit in the
2716 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2717
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2718 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2719 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2720 constant. TYPE must be a scalar or vector floating point type. CST must be
2721 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2722 vectors of the same number of elements. If the value won't fit in the
2723 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2724
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2725 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2726 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2727 constant. TYPE must be a scalar or vector floating point type. CST must be
2728 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2729 vectors of the same number of elements. If the value won't fit in the
2730 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2731
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2732 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2733 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2734 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2735 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2736 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2737
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2738 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2739 <dd>Convert a integer constant to a pointer constant. TYPE must be a pointer
2740 type. CST must be of integer type. The CST value is zero extended,
2741 truncated, or unchanged to make it fit in a pointer size. This one is
2742 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2743
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2744 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +0000[diff] [blame]2745 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2746 are the same as those for the <a href="#i_bitcast">bitcast
2747 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2748
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2749 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2750 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2751 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2752 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2753 instruction, the index list may have zero or more indexes, which are
2754 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2755
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2756 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2757 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2758
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2759 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2760 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2761
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2762 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2763 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2764
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2765 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2766 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2767 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2768
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2769 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2770 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2771 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2772
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2773 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2774 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2775 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2776
Nick Lewycky9e130ce2010-05-29 06:44:15 +0000[diff] [blame]2777 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2778 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2779 constants. The index list is interpreted in a similar manner as indices in
2780 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2781 index value must be specified.</dd>
2782
2783 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2784 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2785 constants. The index list is interpreted in a similar manner as indices in
2786 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2787 index value must be specified.</dd>
2788
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2789 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2790 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2791 be any of the <a href="#binaryops">binary</a>
2792 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2793 on operands are the same as those for the corresponding instruction
2794 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2795</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2796
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2797</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]2798
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2799</div>
2800
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2801<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2802<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2803<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2804<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2805<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2806<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2807<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2808</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2809
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2810<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2811
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2812<p>LLVM supports inline assembler expressions (as opposed
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2813 to <a href="#moduleasm">Module-Level Inline Assembly</a>) through the use of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2814 a special value. This value represents the inline assembler as a string
2815 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2816 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2817 expression has side effects, and a flag indicating whether the function
2818 containing the asm needs to align its stack conservatively. An example
2819 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2820
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2821<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2822i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2823</pre>
2824
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2825<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
2826 a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we
2827 have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2828
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2829<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2830%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]2831</pre>
2832
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2833<p>Inline asms with side effects not visible in the constraint list must be
2834 marked as having side effects. This is done through the use of the
2835 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2836
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2837<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2838call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2839</pre>
2840
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2841<p>In some cases inline asms will contain code that will not work unless the
2842 stack is aligned in some way, such as calls or SSE instructions on x86,
2843 yet will not contain code that does that alignment within the asm.
2844 The compiler should make conservative assumptions about what the asm might
2845 contain and should generate its usual stack alignment code in the prologue
2846 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2847
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2848<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2849call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2850</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2851
2852<p>If both keywords appear the '<tt>sideeffect</tt>' keyword must come
2853 first.</p>
2854
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2855<!--
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2856<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2857 documented here. Constraints on what can be done (e.g. duplication, moving,
2858 etc need to be documented). This is probably best done by reference to
2859 another document that covers inline asm from a holistic perspective.</p>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2860 -->
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2861
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2862<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2863<h4>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2864 <a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2865</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2866
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2867<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2868
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2869<p>The call instructions that wrap inline asm nodes may have a
2870 "<tt>!srcloc</tt>" MDNode attached to it that contains a list of constant
2871 integers. If present, the code generator will use the integer as the
2872 location cookie value when report errors through the <tt>LLVMContext</tt>
2873 error reporting mechanisms. This allows a front-end to correlate backend
2874 errors that occur with inline asm back to the source code that produced it.
2875 For example:</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2876
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2877<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2878call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2879...
2880!42 = !{ i32 1234567 }
2881</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2882
2883<p>It is up to the front-end to make sense of the magic numbers it places in the
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2884 IR. If the MDNode contains multiple constants, the code generator will use
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2885 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]2886
2887</div>
2888
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2889</div>
2890
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2891<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2892<h3>
2893 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2894</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2895
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2896<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2897
2898<p>LLVM IR allows metadata to be attached to instructions in the program that
2899 can convey extra information about the code to the optimizers and code
2900 generator. One example application of metadata is source-level debug
2901 information. There are two metadata primitives: strings and nodes. All
2902 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2903 preceding exclamation point ('<tt>!</tt>').</p>
2904
2905<p>A metadata string is a string surrounded by double quotes. It can contain
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2906 any character by escaping non-printable characters with "<tt>\xx</tt>" where
2907 "<tt>xx</tt>" is the two digit hex code. For example:
2908 "<tt>!"test\00"</tt>".</p>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2909
2910<p>Metadata nodes are represented with notation similar to structure constants
2911 (a comma separated list of elements, surrounded by braces and preceded by an
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2912 exclamation point). Metadata nodes can have any values as their operand. For
2913 example:</p>
2914
2915<div class="doc_code">
2916<pre>
2917!{ metadata !"test\00", i32 10}
2918</pre>
2919</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2920
2921<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2922 metadata nodes, which can be looked up in the module symbol table. For
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2923 example:</p>
2924
2925<div class="doc_code">
2926<pre>
2927!foo = metadata !{!4, !3}
2928</pre>
2929</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2930
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2931<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2932 function is using two metadata arguments:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2933
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2934<div class="doc_code">
2935<pre>
2936call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2937</pre>
2938</div>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2939
2940<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2941 attached to the <tt>add</tt> instruction using the <tt>!dbg</tt>
2942 identifier:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2943
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2944<div class="doc_code">
2945<pre>
2946%indvar.next = add i64 %indvar, 1, !dbg !21
2947</pre>
2948</div>
2949
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]2950<p>More information about specific metadata nodes recognized by the optimizers
2951 and code generator is found below.</p>
2952
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2953<!-- _______________________________________________________________________ -->
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]2954<h4>
2955 <a name="tbaa">'<tt>tbaa</tt>' Metadata</a>
2956</h4>
2957
2958<div>
2959
2960<p>In LLVM IR, memory does not have types, so LLVM's own type system is not
2961 suitable for doing TBAA. Instead, metadata is added to the IR to describe
2962 a type system of a higher level language. This can be used to implement
2963 typical C/C++ TBAA, but it can also be used to implement custom alias
2964 analysis behavior for other languages.</p>
2965
2966<p>The current metadata format is very simple. TBAA metadata nodes have up to
2967 three fields, e.g.:</p>
2968
2969<div class="doc_code">
2970<pre>
2971!0 = metadata !{ metadata !"an example type tree" }
2972!1 = metadata !{ metadata !"int", metadata !0 }
2973!2 = metadata !{ metadata !"float", metadata !0 }
2974!3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
2975</pre>
2976</div>
2977
2978<p>The first field is an identity field. It can be any value, usually
2979 a metadata string, which uniquely identifies the type. The most important
2980 name in the tree is the name of the root node. Two trees with
2981 different root node names are entirely disjoint, even if they
2982 have leaves with common names.</p>
2983
2984<p>The second field identifies the type's parent node in the tree, or
2985 is null or omitted for a root node. A type is considered to alias
2986 all of its descendants and all of its ancestors in the tree. Also,
2987 a type is considered to alias all types in other trees, so that
2988 bitcode produced from multiple front-ends is handled conservatively.</p>
2989
2990<p>If the third field is present, it's an integer which if equal to 1
2991 indicates that the type is "constant" (meaning
2992 <tt>pointsToConstantMemory</tt> should return true; see
2993 <a href="AliasAnalysis.html#OtherItfs">other useful
2994 <tt>AliasAnalysis</tt> methods</a>).</p>
2995
2996</div>
2997
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2998<!-- _______________________________________________________________________ -->
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]2999<h4>
3000 <a name="fpaccuracy">'<tt>fpaccuracy</tt>' Metadata</a>
3001</h4>
3002
3003<div>
3004
3005<p><tt>fpaccuracy</tt> metadata may be attached to any instruction of floating
3006 point type. It expresses the maximum relative error of the result of
3007 that instruction, in ULPs. ULP is defined as follows:</p>
3008
Bill Wendling0656e252011-11-09 19:33:56 +0000[diff] [blame]3009<blockquote>
3010
3011<p>If <tt>x</tt> is a real number that lies between two finite consecutive
3012 floating-point numbers <tt>a</tt> and <tt>b</tt>, without being equal to one
3013 of them, then <tt>ulp(x) = |b - a|</tt>, otherwise <tt>ulp(x)</tt> is the
3014 distance between the two non-equal finite floating-point numbers nearest
3015 <tt>x</tt>. Moreover, <tt>ulp(NaN)</tt> is <tt>NaN</tt>.</p>
3016
3017</blockquote>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3018
3019<p>The maximum relative error may be any rational number. The metadata node
3020 shall consist of a pair of unsigned integers respectively representing
3021 the numerator and denominator. For example, 2.5 ULP:</p>
3022
3023<div class="doc_code">
3024<pre>
3025!0 = metadata !{ i32 5, i32 2 }
3026</pre>
3027</div>
3028
3029</div>
3030
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]3031</div>
3032
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3033</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3034
3035<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3036<h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3037 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3038</h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3039<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3040<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3041<p>LLVM has a number of "magic" global variables that contain data that affect
3042code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3043of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
3044section and all globals that start with "<tt>llvm.</tt>" are reserved for use
3045by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3046
3047<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3048<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3049<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3050</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3051
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3052<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3053
3054<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
3055href="#linkage_appending">appending linkage</a>. This array contains a list of
3056pointers to global variables and functions which may optionally have a pointer
3057cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
3058
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3059<div class="doc_code">
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3060<pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3061@X = global i8 4
3062@Y = global i32 123
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3063
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3064@llvm.used = appending global [2 x i8*] [
3065 i8* @X,
3066 i8* bitcast (i32* @Y to i8*)
3067], section "llvm.metadata"
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3068</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3069</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3070
3071<p>If a global variable appears in the <tt>@llvm.used</tt> list, then the
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3072 compiler, assembler, and linker are required to treat the symbol as if there
3073 is a reference to the global that it cannot see. For example, if a variable
3074 has internal linkage and no references other than that from
3075 the <tt>@llvm.used</tt> list, it cannot be deleted. This is commonly used to
3076 represent references from inline asms and other things the compiler cannot
3077 "see", and corresponds to "<tt>attribute((used))</tt>" in GNU C.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3078
3079<p>On some targets, the code generator must emit a directive to the assembler or
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3080 object file to prevent the assembler and linker from molesting the
3081 symbol.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3082
3083</div>
3084
3085<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3086<h3>
3087 <a name="intg_compiler_used">
3088 The '<tt>llvm.compiler.used</tt>' Global Variable
3089 </a>
3090</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3091
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3092<div>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3093
3094<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3095 <tt>@llvm.used</tt> directive, except that it only prevents the compiler from
3096 touching the symbol. On targets that support it, this allows an intelligent
3097 linker to optimize references to the symbol without being impeded as it would
3098 be by <tt>@llvm.used</tt>.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3099
3100<p>This is a rare construct that should only be used in rare circumstances, and
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3101 should not be exposed to source languages.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3102
3103</div>
3104
3105<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3106<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3107<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3108</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3109
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3110<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3111
3112<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3113<pre>
3114%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3115@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3116</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3117</div>
3118
3119<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor
3120 functions and associated priorities. The functions referenced by this array
3121 will be called in ascending order of priority (i.e. lowest first) when the
3122 module is loaded. The order of functions with the same priority is not
3123 defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3124
3125</div>
3126
3127<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3128<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3129<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3130</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3131
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3132<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3133
3134<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3135<pre>
3136%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3137@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3138</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3139</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3140
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3141<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions
3142 and associated priorities. The functions referenced by this array will be
3143 called in descending order of priority (i.e. highest first) when the module
3144 is loaded. The order of functions with the same priority is not defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3145
3146</div>
3147
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3148</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3149
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]3150<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3151<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3152<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3153
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3154<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3155
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3156<p>The LLVM instruction set consists of several different classifications of
3157 instructions: <a href="#terminators">terminator
3158 instructions</a>, <a href="#binaryops">binary instructions</a>,
3159 <a href="#bitwiseops">bitwise binary instructions</a>,
3160 <a href="#memoryops">memory instructions</a>, and
3161 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3162
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3163<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3164<h3>
3165 <a name="terminators">Terminator Instructions</a>
3166</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3167
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3168<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3169
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3170<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3171 in a program ends with a "Terminator" instruction, which indicates which
3172 block should be executed after the current block is finished. These
3173 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3174 control flow, not values (the one exception being the
3175 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3176
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3177<p>The terminator instructions are:
3178 '<a href="#i_ret"><tt>ret</tt></a>',
3179 '<a href="#i_br"><tt>br</tt></a>',
3180 '<a href="#i_switch"><tt>switch</tt></a>',
3181 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>',
3182 '<a href="#i_invoke"><tt>invoke</tt></a>',
3183 '<a href="#i_unwind"><tt>unwind</tt></a>',
3184 '<a href="#i_resume"><tt>resume</tt></a>', and
3185 '<a href="#i_unreachable"><tt>unreachable</tt></a>'.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3186
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3187<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3188<h4>
3189 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3190</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3191
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3192<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3193
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3194<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3195<pre>
3196 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]3197 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3198</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3199
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3200<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3201<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3202 a value) from a function back to the caller.</p>
3203
3204<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3205 value and then causes control flow, and one that just causes control flow to
3206 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3207
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3208<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3209<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3210 return value. The type of the return value must be a
3211 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3212
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3213<p>A function is not <a href="#wellformed">well formed</a> if it it has a
3214 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3215 value or a return value with a type that does not match its type, or if it
3216 has a void return type and contains a '<tt>ret</tt>' instruction with a
3217 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3218
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3219<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3220<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3221 the calling function's context. If the caller is a
3222 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3223 instruction after the call. If the caller was an
3224 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3225 the beginning of the "normal" destination block. If the instruction returns
3226 a value, that value shall set the call or invoke instruction's return
3227 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3228
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3229<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3230<pre>
3231 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]3232 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +0000[diff] [blame]3233 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]3234</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]3235
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3236</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3237<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3238<h4>
3239 <a name="i_br">'<tt>br</tt>' Instruction</a>
3240</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3241
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3242<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3243
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3244<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3245<pre>
Bill Wendlingb3aa4712011-07-26 10:41:15 +0000[diff] [blame]3246 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3247 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3248</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3249
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3250<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3251<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3252 different basic block in the current function. There are two forms of this
3253 instruction, corresponding to a conditional branch and an unconditional
3254 branch.</p>
3255
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3256<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3257<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3258 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3259 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3260 target.</p>
3261
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3262<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3263<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3264 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3265 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3266 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3267
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3268<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]3269<pre>
3270Test:
3271 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3272 br i1 %cond, label %IfEqual, label %IfUnequal
3273IfEqual:
3274 <a href="#i_ret">ret</a> i32 1
3275IfUnequal:
3276 <a href="#i_ret">ret</a> i32 0
3277</pre>
3278
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3279</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3280
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3281<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3282<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3283 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3284</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3285
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3286<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3287
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3288<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3289<pre>
3290 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3291</pre>
3292
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3293<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3294<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]3295 several different places. It is a generalization of the '<tt>br</tt>'
3296 instruction, allowing a branch to occur to one of many possible
3297 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3298
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3299<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3300<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3301 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3302 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3303 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3304
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3305<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3306<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3307 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3308 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3309 transferred to the corresponding destination; otherwise, control flow is
3310 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3311
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3312<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3313<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3314 <tt>switch</tt> instruction, this instruction may be code generated in
3315 different ways. For example, it could be generated as a series of chained
3316 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3317
3318<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3319<pre>
3320 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3321 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3322 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3323
3324 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3325 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3326
3327 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3328 switch i32 %val, label %otherwise [ i32 0, label %onzero
3329 i32 1, label %onone
3330 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3331</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3332
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3333</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3334
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3335
3336<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3337<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3338 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3339</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3340
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3341<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3342
3343<h5>Syntax:</h5>
3344<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3345 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]3346</pre>
3347
3348<h5>Overview:</h5>
3349
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3350<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3351 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]3352 "<tt>address</tt>". Address must be derived from a <a
3353 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3354
3355<h5>Arguments:</h5>
3356
3357<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3358 rest of the arguments indicate the full set of possible destinations that the
3359 address may point to. Blocks are allowed to occur multiple times in the
3360 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3361
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3362<p>This destination list is required so that dataflow analysis has an accurate
3363 understanding of the CFG.</p>
3364
3365<h5>Semantics:</h5>
3366
3367<p>Control transfers to the block specified in the address argument. All
3368 possible destination blocks must be listed in the label list, otherwise this
3369 instruction has undefined behavior. This implies that jumps to labels
3370 defined in other functions have undefined behavior as well.</p>
3371
3372<h5>Implementation:</h5>
3373
3374<p>This is typically implemented with a jump through a register.</p>
3375
3376<h5>Example:</h5>
3377<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3378 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3379</pre>
3380
3381</div>
3382
3383
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3384<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3385<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3386 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3387</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3388
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3389<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3390
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3391<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3392<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3393 &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]3394 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3395</pre>
3396
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3397<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3398<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3399 function, with the possibility of control flow transfer to either the
3400 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3401 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3402 control flow will return to the "normal" label. If the callee (or any
3403 indirect callees) returns with the "<a href="#i_unwind"><tt>unwind</tt></a>"
3404 instruction, control is interrupted and continued at the dynamically nearest
3405 "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3406
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3407<p>The '<tt>exception</tt>' label is a
3408 <i><a href="ExceptionHandling.html#overview">landing pad</a></i> for the
3409 exception. As such, '<tt>exception</tt>' label is required to have the
3410 "<a href="#i_landingpad"><tt>landingpad</tt></a>" instruction, which contains
Chad Rosier85f5a1a2011-12-09 02:00:44 +0000[diff] [blame]3411 the information about the behavior of the program after unwinding
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3412 happens, as its first non-PHI instruction. The restrictions on the
3413 "<tt>landingpad</tt>" instruction's tightly couples it to the
3414 "<tt>invoke</tt>" instruction, so that the important information contained
3415 within the "<tt>landingpad</tt>" instruction can't be lost through normal
3416 code motion.</p>
3417
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3418<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3419<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3420
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3421<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3422 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3423 convention</a> the call should use. If none is specified, the call
3424 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3425
3426 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3427 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3428 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3429
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3430 <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]3431 function value being invoked. In most cases, this is a direct function
3432 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3433 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3434
3435 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3436 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3437
3438 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]3439 signature argument types and parameter attributes. All arguments must be
3440 of <a href="#t_firstclass">first class</a> type. If the function
3441 signature indicates the function accepts a variable number of arguments,
3442 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3443
3444 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3445 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3446
3447 <li>'<tt>exception label</tt>': the label reached when a callee returns with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3448 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3449
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3450 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3451 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3452 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3453</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3454
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3455<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3456<p>This instruction is designed to operate as a standard
3457 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3458 primary difference is that it establishes an association with a label, which
3459 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3460
3461<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3462 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3463 exception. Additionally, this is important for implementation of
3464 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3465
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3466<p>For the purposes of the SSA form, the definition of the value returned by the
3467 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3468 block to the "normal" label. If the callee unwinds then no return value is
3469 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +0000[diff] [blame]3470
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3471<p>Note that the code generator does not yet completely support unwind, and
3472that the invoke/unwind semantics are likely to change in future versions.</p>
3473
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3474<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3475<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3476 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3477 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3478 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3479 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3480</pre>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3481
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3482</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3483
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3484<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3485
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3486<h4>
3487 <a name="i_unwind">'<tt>unwind</tt>' Instruction</a>
3488</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3489
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3490<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3491
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3492<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3493<pre>
3494 unwind
3495</pre>
3496
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3497<h5>Overview:</h5>
Bill Wendling5a658642012-02-06 17:58:34 +0000[diff] [blame]3498<p><span class="doc_warning">N.B. The <tt>unwind</tt> instruction has been
3499 deprecated and is slated for removal.</span></p>
3500
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3501<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3502 at the first callee in the dynamic call stack which used
3503 an <a href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call.
3504 This is primarily used to implement exception handling.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3505
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]3506<h5>Semantics:</h5>
Chris Lattner72ed2002008-04-19 21:01:16 +0000[diff] [blame]3507<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3508 immediately halt. The dynamic call stack is then searched for the
3509 first <a href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack.
3510 Once found, execution continues at the "exceptional" destination block
3511 specified by the <tt>invoke</tt> instruction. If there is no <tt>invoke</tt>
3512 instruction in the dynamic call chain, undefined behavior results.</p>
3513
Chris Lattnerdf7a6802010-01-15 18:08:37 +0000[diff] [blame]3514<p>Note that the code generator does not yet completely support unwind, and
3515that the invoke/unwind semantics are likely to change in future versions.</p>
3516
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3517</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3518
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3519 <!-- _______________________________________________________________________ -->
3520
3521<h4>
3522 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3523</h4>
3524
3525<div>
3526
3527<h5>Syntax:</h5>
3528<pre>
3529 resume &lt;type&gt; &lt;value&gt;
3530</pre>
3531
3532<h5>Overview:</h5>
3533<p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3534 successors.</p>
3535
3536<h5>Arguments:</h5>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3537<p>The '<tt>resume</tt>' instruction requires one argument, which must have the
Bill Wendlinge4ad50b2011-08-03 18:37:32 +0000[diff] [blame]3538 same type as the result of any '<tt>landingpad</tt>' instruction in the same
3539 function.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3540
3541<h5>Semantics:</h5>
3542<p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3543 (in-flight) exception whose unwinding was interrupted with
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3544 a <a href="#i_landingpad"><tt>landingpad</tt></a> instruction.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3545
3546<h5>Example:</h5>
3547<pre>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3548 resume { i8*, i32 } %exn
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3549</pre>
3550
3551</div>
3552
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3553<!-- _______________________________________________________________________ -->
3554
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3555<h4>
3556 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3557</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3558
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3559<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3560
3561<h5>Syntax:</h5>
3562<pre>
3563 unreachable
3564</pre>
3565
3566<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3567<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3568 instruction is used to inform the optimizer that a particular portion of the
3569 code is not reachable. This can be used to indicate that the code after a
3570 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3571
3572<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3573<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3574
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3575</div>
3576
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3577</div>
3578
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3579<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3580<h3>
3581 <a name="binaryops">Binary Operations</a>
3582</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3583
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3584<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3585
3586<p>Binary operators are used to do most of the computation in a program. They
3587 require two operands of the same type, execute an operation on them, and
3588 produce a single value. The operands might represent multiple data, as is
3589 the case with the <a href="#t_vector">vector</a> data type. The result value
3590 has the same type as its operands.</p>
3591
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3592<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3593
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3594<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3595<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3596 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3597</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3598
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3599<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3600
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3601<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3602<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3603 &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]3604 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3605 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3606 &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]3607</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3608
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3609<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3610<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3611
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3612<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3613<p>The two arguments to the '<tt>add</tt>' instruction must
3614 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3615 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3616
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3617<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3618<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3619
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3620<p>If the sum has unsigned overflow, the result returned is the mathematical
3621 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]3622
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3623<p>Because LLVM integers use a two's complement representation, this instruction
3624 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3625
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3626<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3627 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3628 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3629 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3630 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3631
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3632<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3633<pre>
3634 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3635</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3636
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3637</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3638
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3639<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3640<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3641 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3642</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3643
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3644<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3645
3646<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3647<pre>
3648 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3649</pre>
3650
3651<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3652<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3653
3654<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3655<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3656 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3657 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3658
3659<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3660<p>The value produced is the floating point sum of the two operands.</p>
3661
3662<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3663<pre>
3664 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3665</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3666
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3667</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3668
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3669<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3670<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3671 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3672</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3673
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3674<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3675
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3676<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3677<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3678 &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]3679 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3680 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3681 &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]3682</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3683
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3684<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3685<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3686 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3687
3688<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3689 '<tt>neg</tt>' instruction present in most other intermediate
3690 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3691
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3692<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3693<p>The two arguments to the '<tt>sub</tt>' instruction must
3694 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3695 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3696
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3697<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3698<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3699
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3700<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3701 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3702 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3703
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3704<p>Because LLVM integers use a two's complement representation, this instruction
3705 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3706
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3707<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3708 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3709 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3710 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3711 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3712
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3713<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]3714<pre>
3715 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3716 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3717</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3718
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3719</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3720
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3721<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3722<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3723 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3724</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3725
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3726<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3727
3728<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3729<pre>
3730 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3731</pre>
3732
3733<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3734<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3735 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3736
3737<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3738 '<tt>fneg</tt>' instruction present in most other intermediate
3739 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3740
3741<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]3742<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3743 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3744 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3745
3746<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3747<p>The value produced is the floating point difference of the two operands.</p>
3748
3749<h5>Example:</h5>
3750<pre>
3751 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
3752 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
3753</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3754
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3755</div>
3756
3757<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3758<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3759 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3760</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3761
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3762<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3763
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3764<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3765<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3766 &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]3767 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3768 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3769 &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]3770</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3771
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3772<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3773<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3774
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3775<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3776<p>The two arguments to the '<tt>mul</tt>' instruction must
3777 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3778 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3779
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3780<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3781<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3782
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3783<p>If the result of the multiplication has unsigned overflow, the result
3784 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
3785 width of the result.</p>
3786
3787<p>Because LLVM integers use a two's complement representation, and the result
3788 is the same width as the operands, this instruction returns the correct
3789 result for both signed and unsigned integers. If a full product
3790 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
3791 be sign-extended or zero-extended as appropriate to the width of the full
3792 product.</p>
3793
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3794<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3795 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3796 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3797 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3798 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3799
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3800<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3801<pre>
3802 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3803</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3804
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3805</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3806
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3807<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3808<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3809 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3810</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3811
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3812<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3813
3814<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3815<pre>
3816 &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]3817</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3818
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3819<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3820<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3821
3822<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3823<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3824 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3825 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3826
3827<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3828<p>The value produced is the floating point product of the two operands.</p>
3829
3830<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3831<pre>
3832 &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]3833</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3834
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3835</div>
3836
3837<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3838<h4>
3839 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
3840</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3841
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3842<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3843
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3844<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3845<pre>
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3846 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3847 &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]3848</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3849
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3850<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3851<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3852
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3853<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3854<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3855 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3856 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3857
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3858<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3859<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3860
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3861<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3862 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
3863
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3864<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3865
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3866<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3867 <tt>udiv</tt> is a <a href="#poisonvalues">poison value</a> if %op1 is not a
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]3868 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
3869
3870
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3871<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3872<pre>
3873 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3874</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3875
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3876</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3877
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3878<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3879<h4>
3880 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
3881</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3882
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3883<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3884
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3885<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3886<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3887 &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]3888 &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]3889</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3890
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3891<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3892<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3893
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3894<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3895<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3896 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
3897 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3898
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3899<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3900<p>The value produced is the signed integer quotient of the two operands rounded
3901 towards zero.</p>
3902
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3903<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3904 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
3905
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3906<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3907 undefined behavior; this is a rare case, but can occur, for example, by doing
3908 a 32-bit division of -2147483648 by -1.</p>
3909
Dan Gohman9c5beed2009-07-22 00:04:19 +0000[diff] [blame]3910<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3911 <tt>sdiv</tt> is a <a href="#poisonvalues">poison value</a> if the result would
Dan Gohman38da9272010-07-11 00:08:34 +0000[diff] [blame]3912 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3913
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3914<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3915<pre>
3916 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3917</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3918
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3919</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3920
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3921<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3922<h4>
3923 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
3924</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3925
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3926<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3927
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3928<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3929<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3930 &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]3931</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3932
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3933<h5>Overview:</h5>
3934<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3935
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3936<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3937<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3938 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3939 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3940
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3941<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]3942<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3943
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3944<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3945<pre>
3946 &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]3947</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3948
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3949</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3950
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3951<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3952<h4>
3953 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
3954</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3955
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3956<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3957
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3958<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3959<pre>
3960 &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]3961</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3962
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3963<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3964<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
3965 division of its two arguments.</p>
3966
Reid Spencer0a783f72006-11-02 01:53:59 +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>urem</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>
3971
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3972<h5>Semantics:</h5>
3973<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3974 This instruction always performs an unsigned division to get the
3975 remainder.</p>
3976
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3977<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3978 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
3979
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]3980<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3981
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3982<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3983<pre>
3984 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3985</pre>
3986
3987</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3988
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]3989<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3990<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3991 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3992</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3993
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3994<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3995
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3996<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3997<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]3998 &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]3999</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4000
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4001<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4002<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
4003 division of its two operands. This instruction can also take
4004 <a href="#t_vector">vector</a> versions of the values in which case the
4005 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]4006
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4007<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4008<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4009 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4010 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4011
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4012<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4013<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +0000[diff] [blame]4014 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
4015 <i>modulo</i> operator (where the result is either zero or has the same sign
4016 as the divisor, <tt>op2</tt>) of a value.
4017 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4018 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
4019 Math Forum</a>. For a table of how this is implemented in various languages,
4020 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
4021 Wikipedia: modulo operation</a>.</p>
4022
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4023<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4024 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
4025
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4026<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4027 Overflow also leads to undefined behavior; this is a rare case, but can
4028 occur, for example, by taking the remainder of a 32-bit division of
4029 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
4030 lets srem be implemented using instructions that return both the result of
4031 the division and the remainder.)</p>
4032
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4033<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4034<pre>
4035 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4036</pre>
4037
4038</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4039
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4040<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4041<h4>
4042 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
4043</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4044
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4045<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4046
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4047<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4048<pre>
4049 &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]4050</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4051
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4052<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4053<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
4054 its two operands.</p>
4055
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4056<h5>Arguments:</h5>
4057<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4058 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4059 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4060
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4061<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4062<p>This instruction returns the <i>remainder</i> of a division. The remainder
4063 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4064
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4065<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4066<pre>
4067 &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]4068</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4069
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4070</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]4071
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4072</div>
4073
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4074<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4075<h3>
4076 <a name="bitwiseops">Bitwise Binary Operations</a>
4077</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4078
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4079<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4080
4081<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
4082 program. They are generally very efficient instructions and can commonly be
4083 strength reduced from other instructions. They require two operands of the
4084 same type, execute an operation on them, and produce a single value. The
4085 resulting value is the same type as its operands.</p>
4086
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4087<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4088<h4>
4089 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
4090</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4091
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4092<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4093
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4094<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4095<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4096 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4097 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4098 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4099 &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]4100</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4101
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4102<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4103<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
4104 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4105
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4106<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4107<p>Both arguments to the '<tt>shl</tt>' instruction must be the
4108 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
4109 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4110
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4111<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4112<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
4113 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
4114 is (statically or dynamically) negative or equal to or larger than the number
4115 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4116 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4117 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4118
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4119<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4120 <a href="#poisonvalues">poison value</a> if it shifts out any non-zero bits. If
Chris Lattner66298c12011-02-09 16:44:44 +0000[diff] [blame]4121 the <tt>nsw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4122 <a href="#poisonvalues">poison value</a> if it shifts out any bits that disagree
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4123 with the resultant sign bit. As such, NUW/NSW have the same semantics as
4124 they would if the shift were expressed as a mul instruction with the same
4125 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
4126
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4127<h5>Example:</h5>
4128<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4129 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
4130 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
4131 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4132 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4133 &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]4134</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4135
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4136</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4137
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4138<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4139<h4>
4140 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
4141</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4142
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4143<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4144
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4145<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4146<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4147 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4148 &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]4149</pre>
4150
4151<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4152<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
4153 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4154
4155<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4156<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4157 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4158 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4159
4160<h5>Semantics:</h5>
4161<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4162 significant bits of the result will be filled with zero bits after the shift.
4163 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4164 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4165 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4166 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4167
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4168<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4169 <tt>lshr</tt> is a <a href="#poisonvalues">poison value</a> if any of the bits
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4170 shifted out are non-zero.</p>
4171
4172
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4173<h5>Example:</h5>
4174<pre>
4175 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4176 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4177 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4178 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4179 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4180 &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]4181</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4182
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4183</div>
4184
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4185<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4186<h4>
4187 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4188</h4>
4189
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4190<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4191
4192<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4193<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4194 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4195 &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]4196</pre>
4197
4198<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4199<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4200 operand shifted to the right a specified number of bits with sign
4201 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4202
4203<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4204<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4205 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4206 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4207
4208<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4209<p>This instruction always performs an arithmetic shift right operation, The
4210 most significant bits of the result will be filled with the sign bit
4211 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4212 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4213 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4214 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4215
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4216<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4217 <tt>ashr</tt> is a <a href="#poisonvalues">poison value</a> if any of the bits
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4218 shifted out are non-zero.</p>
4219
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4220<h5>Example:</h5>
4221<pre>
4222 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4223 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4224 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4225 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4226 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4227 &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]4228</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4229
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4230</div>
4231
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4232<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4233<h4>
4234 <a name="i_and">'<tt>and</tt>' Instruction</a>
4235</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4236
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4237<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4238
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4239<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4240<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4241 &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]4242</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4243
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4244<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4245<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4246 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4247
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4248<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4249<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4250 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4251 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4252
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4253<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4254<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4255
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4256<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4257 <tbody>
4258 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4259 <th>In0</th>
4260 <th>In1</th>
4261 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4262 </tr>
4263 <tr>
4264 <td>0</td>
4265 <td>0</td>
4266 <td>0</td>
4267 </tr>
4268 <tr>
4269 <td>0</td>
4270 <td>1</td>
4271 <td>0</td>
4272 </tr>
4273 <tr>
4274 <td>1</td>
4275 <td>0</td>
4276 <td>0</td>
4277 </tr>
4278 <tr>
4279 <td>1</td>
4280 <td>1</td>
4281 <td>1</td>
4282 </tr>
4283 </tbody>
4284</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4285
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4286<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4287<pre>
4288 &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]4289 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4290 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4291</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4292</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4293<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4294<h4>
4295 <a name="i_or">'<tt>or</tt>' Instruction</a>
4296</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4297
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4298<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4299
4300<h5>Syntax:</h5>
4301<pre>
4302 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4303</pre>
4304
4305<h5>Overview:</h5>
4306<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4307 two operands.</p>
4308
4309<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4310<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4311 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4312 values. Both arguments must have identical types.</p>
4313
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4314<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4315<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4316
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4317<table border="1" cellspacing="0" cellpadding="4">
4318 <tbody>
4319 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4320 <th>In0</th>
4321 <th>In1</th>
4322 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4323 </tr>
4324 <tr>
4325 <td>0</td>
4326 <td>0</td>
4327 <td>0</td>
4328 </tr>
4329 <tr>
4330 <td>0</td>
4331 <td>1</td>
4332 <td>1</td>
4333 </tr>
4334 <tr>
4335 <td>1</td>
4336 <td>0</td>
4337 <td>1</td>
4338 </tr>
4339 <tr>
4340 <td>1</td>
4341 <td>1</td>
4342 <td>1</td>
4343 </tr>
4344 </tbody>
4345</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4346
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4347<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4348<pre>
4349 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4350 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4351 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4352</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4353
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4354</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4355
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4356<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4357<h4>
4358 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4359</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4360
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4361<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4362
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4363<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4364<pre>
4365 &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]4366</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4367
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4368<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4369<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4370 its two operands. The <tt>xor</tt> is used to implement the "one's
4371 complement" operation, which is the "~" operator in C.</p>
4372
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4373<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4374<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4375 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4376 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4377
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4378<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4379<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4380
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4381<table border="1" cellspacing="0" cellpadding="4">
4382 <tbody>
4383 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4384 <th>In0</th>
4385 <th>In1</th>
4386 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4387 </tr>
4388 <tr>
4389 <td>0</td>
4390 <td>0</td>
4391 <td>0</td>
4392 </tr>
4393 <tr>
4394 <td>0</td>
4395 <td>1</td>
4396 <td>1</td>
4397 </tr>
4398 <tr>
4399 <td>1</td>
4400 <td>0</td>
4401 <td>1</td>
4402 </tr>
4403 <tr>
4404 <td>1</td>
4405 <td>1</td>
4406 <td>0</td>
4407 </tr>
4408 </tbody>
4409</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4410
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4411<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4412<pre>
4413 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4414 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4415 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4416 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4417</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4418
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4419</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4420
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4421</div>
4422
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4423<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4424<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4425 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4426</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4427
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4428<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4429
4430<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4431 target-independent manner. These instructions cover the element-access and
4432 vector-specific operations needed to process vectors effectively. While LLVM
4433 does directly support these vector operations, many sophisticated algorithms
4434 will want to use target-specific intrinsics to take full advantage of a
4435 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4436
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4437<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4438<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4439 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4440</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4441
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4442<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4443
4444<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4445<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4446 &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]4447</pre>
4448
4449<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4450<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4451 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4452
4453
4454<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4455<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4456 of <a href="#t_vector">vector</a> type. The second operand is an index
4457 indicating the position from which to extract the element. The index may be
4458 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4459
4460<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4461<p>The result is a scalar of the same type as the element type of
4462 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4463 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4464 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4465
4466<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4467<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4468 &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]4469</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4470
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4471</div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4472
4473<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4474<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4475 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4476</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4477
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4478<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4479
4480<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4481<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +0000[diff] [blame]4482 &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]4483</pre>
4484
4485<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4486<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4487 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4488
4489<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4490<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4491 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4492 whose type must equal the element type of the first operand. The third
4493 operand is an index indicating the position at which to insert the value.
4494 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4495
4496<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4497<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4498 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4499 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4500 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4501
4502<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4503<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4504 &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]4505</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4506
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4507</div>
4508
4509<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4510<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4511 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4512</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4513
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4514<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4515
4516<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4517<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4518 &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]4519</pre>
4520
4521<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4522<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4523 from two input vectors, returning a vector with the same element type as the
4524 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4525
4526<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4527<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
4528 with types that match each other. The third argument is a shuffle mask whose
4529 element type is always 'i32'. The result of the instruction is a vector
4530 whose length is the same as the shuffle mask and whose element type is the
4531 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4532
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4533<p>The shuffle mask operand is required to be a constant vector with either
4534 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4535
4536<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4537<p>The elements of the two input vectors are numbered from left to right across
4538 both of the vectors. The shuffle mask operand specifies, for each element of
4539 the result vector, which element of the two input vectors the result element
4540 gets. The element selector may be undef (meaning "don't care") and the
4541 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4542
4543<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4544<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4545 &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]4546 &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]4547 &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]4548 &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]4549 &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]4550 &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]4551 &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]4552 &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]4553</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4554
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4555</div>
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]4556
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4557</div>
4558
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4559<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4560<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4561 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4562</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4563
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4564<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4565
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4566<p>LLVM supports several instructions for working with
4567 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4568
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4569<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4570<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4571 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4572</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4573
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4574<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4575
4576<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4577<pre>
4578 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4579</pre>
4580
4581<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4582<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4583 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4584
4585<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4586<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4587 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4588 <a href="#t_array">array</a> type. The operands are constant indices to
4589 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4590 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4591 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4592 <ul>
4593 <li>Since the value being indexed is not a pointer, the first index is
4594 omitted and assumed to be zero.</li>
4595 <li>At least one index must be specified.</li>
4596 <li>Not only struct indices but also array indices must be in
4597 bounds.</li>
4598 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4599
4600<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4601<p>The result is the value at the position in the aggregate specified by the
4602 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4603
4604<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4605<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4606 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4607</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4608
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4609</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4610
4611<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4612<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4613 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4614</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4615
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4616<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4617
4618<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4619<pre>
Bill Wendling194229e2011-07-26 20:42:28 +0000[diff] [blame]4620 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, &lt;idx&gt;{, &lt;idx&gt;}* <i>; yields &lt;aggregate type&gt;</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4621</pre>
4622
4623<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4624<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4625 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4626
4627<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4628<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4629 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4630 <a href="#t_array">array</a> type. The second operand is a first-class
4631 value to insert. The following operands are constant indices indicating
4632 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]4633 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4634 value to insert must have the same type as the value identified by the
4635 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4636
4637<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4638<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4639 that of <tt>val</tt> except that the value at the position specified by the
4640 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4641
4642<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4643<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +0000[diff] [blame]4644 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4645 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4646 %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]4647</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4648
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4649</div>
4650
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4651</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4652
4653<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4654<h3>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4655 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4656</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4657
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4658<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4659
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4660<p>A key design point of an SSA-based representation is how it represents
4661 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]4662 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4663 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4664
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4665<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4666<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4667 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4668</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4669
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4670<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4671
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4672<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4673<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +0000[diff] [blame]4674 &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]4675</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4676
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4677<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4678<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4679 currently executing function, to be automatically released when this function
4680 returns to its caller. The object is always allocated in the generic address
4681 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4682
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4683<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4684<p>The '<tt>alloca</tt>' instruction
4685 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4686 runtime stack, returning a pointer of the appropriate type to the program.
4687 If "NumElements" is specified, it is the number of elements allocated,
4688 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4689 specified, the value result of the allocation is guaranteed to be aligned to
4690 at least that boundary. If not specified, or if zero, the target can choose
4691 to align the allocation on any convenient boundary compatible with the
4692 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4693
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4694<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4695
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4696<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +0000[diff] [blame]4697<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4698 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4699 memory is automatically released when the function returns. The
4700 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4701 variables that must have an address available. When the function returns
4702 (either with the <tt><a href="#i_ret">ret</a></tt>
4703 or <tt><a href="#i_unwind">unwind</a></tt> instructions), the memory is
4704 reclaimed. Allocating zero bytes is legal, but the result is undefined.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4705
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4706<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4707<pre>
Dan Gohman81e21672009-01-04 23:49:44 +0000[diff] [blame]4708 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4709 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4710 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4711 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4712</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4713
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4714</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4715
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4716<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4717<h4>
4718 <a name="i_load">'<tt>load</tt>' Instruction</a>
4719</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4720
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4721<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4722
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4723<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4724<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]4725 &lt;result&gt; = load [volatile] &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;]
4726 &lt;result&gt; = load atomic [volatile] &lt;ty&gt;* &lt;pointer&gt; [singlethread] &lt;ordering&gt;, align &lt;alignment&gt;
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4727 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4728</pre>
4729
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4730<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4731<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4732
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4733<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4734<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4735 from which to load. The pointer must point to
4736 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
4737 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4738 number or order of execution of this <tt>load</tt> with other <a
4739 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4740
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]4741<p>If the <code>load</code> is marked as <code>atomic</code>, it takes an extra
4742 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4743 argument. The <code>release</code> and <code>acq_rel</code> orderings are
4744 not valid on <code>load</code> instructions. Atomic loads produce <a
4745 href="#memorymodel">defined</a> results when they may see multiple atomic
4746 stores. The type of the pointee must be an integer type whose bit width
4747 is a power of two greater than or equal to eight and less than or equal
4748 to a target-specific size limit. <code>align</code> must be explicitly
4749 specified on atomic loads, and the load has undefined behavior if the
4750 alignment is not set to a value which is at least the size in bytes of
4751 the pointee. <code>!nontemporal</code> does not have any defined semantics
4752 for atomic loads.</p>
4753
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4754<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4755 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]4756 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4757 alignment for the target. It is the responsibility of the code emitter to
4758 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4759 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4760 produce less efficient code. An alignment of 1 is always safe.</p>
4761
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4762<p>The optional <tt>!nontemporal</tt> metadata must reference a single
4763 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4764 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]4765 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
4766 and code generator that this load is not expected to be reused in the cache.
4767 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4768 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4769
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4770<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4771<p>The location of memory pointed to is loaded. If the value being loaded is of
4772 scalar type then the number of bytes read does not exceed the minimum number
4773 of bytes needed to hold all bits of the type. For example, loading an
4774 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
4775 <tt>i20</tt> with a size that is not an integral number of bytes, the result
4776 is undefined if the value was not originally written using a store of the
4777 same type.</p>
4778
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4779<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4780<pre>
4781 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
4782 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4783 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4784</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4785
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4786</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4787
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4788<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4789<h4>
4790 <a name="i_store">'<tt>store</tt>' Instruction</a>
4791</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4792
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4793<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4794
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4795<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4796<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4797 store [volatile] &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;] <i>; yields {void}</i>
4798 store atomic [volatile] &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt; [singlethread] &lt;ordering&gt;, align &lt;alignment&gt; <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4799</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4800
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4801<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4802<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4803
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4804<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4805<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
4806 and an address at which to store it. The type of the
4807 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
4808 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]4809 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
4810 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
4811 order of execution of this <tt>store</tt> with other <a
4812 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4813
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]4814<p>If the <code>store</code> is marked as <code>atomic</code>, it takes an extra
4815 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
4816 argument. The <code>acquire</code> and <code>acq_rel</code> orderings aren't
4817 valid on <code>store</code> instructions. Atomic loads produce <a
4818 href="#memorymodel">defined</a> results when they may see multiple atomic
4819 stores. The type of the pointee must be an integer type whose bit width
4820 is a power of two greater than or equal to eight and less than or equal
4821 to a target-specific size limit. <code>align</code> must be explicitly
4822 specified on atomic stores, and the store has undefined behavior if the
4823 alignment is not set to a value which is at least the size in bytes of
4824 the pointee. <code>!nontemporal</code> does not have any defined semantics
4825 for atomic stores.</p>
4826
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4827<p>The optional constant "align" argument specifies the alignment of the
4828 operation (that is, the alignment of the memory address). A value of 0 or an
4829 omitted "align" argument means that the operation has the preferential
4830 alignment for the target. It is the responsibility of the code emitter to
4831 ensure that the alignment information is correct. Overestimating the
4832 alignment results in an undefined behavior. Underestimating the alignment may
4833 produce less efficient code. An alignment of 1 is always safe.</p>
4834
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4835<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]4836 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4837 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4838 instruction tells the optimizer and code generator that this load is
4839 not expected to be reused in the cache. The code generator may
4840 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]4841 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]4842
4843
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4844<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4845<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
4846 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
4847 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
4848 does not exceed the minimum number of bytes needed to hold all bits of the
4849 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
4850 writing a value of a type like <tt>i20</tt> with a size that is not an
4851 integral number of bytes, it is unspecified what happens to the extra bits
4852 that do not belong to the type, but they will typically be overwritten.</p>
4853
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4854<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4855<pre>
4856 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]4857 store i32 3, i32* %ptr <i>; yields {void}</i>
4858 %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]4859</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4860
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]4861</div>
4862
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4863<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4864<h4>
4865<a name="i_fence">'<tt>fence</tt>' Instruction</a>
4866</h4>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]4867
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4868<div>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]4869
4870<h5>Syntax:</h5>
4871<pre>
4872 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
4873</pre>
4874
4875<h5>Overview:</h5>
4876<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
4877between operations.</p>
4878
4879<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
4880href="#ordering">ordering</a> argument which defines what
4881<i>synchronizes-with</i> edges they add. They can only be given
4882<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
4883<code>seq_cst</code> orderings.</p>
4884
4885<h5>Semantics:</h5>
4886<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
4887semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
4888<code>acquire</code> ordering semantics if and only if there exist atomic
4889operations <var>X</var> and <var>Y</var>, both operating on some atomic object
4890<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
4891<var>X</var> modifies <var>M</var> (either directly or through some side effect
4892of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
4893<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
4894<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
4895than an explicit <code>fence</code>, one (but not both) of the atomic operations
4896<var>X</var> or <var>Y</var> might provide a <code>release</code> or
4897<code>acquire</code> (resp.) ordering constraint and still
4898<i>synchronize-with</i> the explicit <code>fence</code> and establish the
4899<i>happens-before</i> edge.</p>
4900
4901<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
4902having both <code>acquire</code> and <code>release</code> semantics specified
4903above, participates in the global program order of other <code>seq_cst</code>
4904operations and/or fences.</p>
4905
4906<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
4907specifies that the fence only synchronizes with other fences in the same
4908thread. (This is useful for interacting with signal handlers.)</p>
4909
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]4910<h5>Example:</h5>
4911<pre>
4912 fence acquire <i>; yields {void}</i>
4913 fence singlethread seq_cst <i>; yields {void}</i>
4914</pre>
4915
4916</div>
4917
4918<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4919<h4>
4920<a name="i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a>
4921</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4922
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]4923<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4924
4925<h5>Syntax:</h5>
4926<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4927 cmpxchg [volatile] &lt;ty&gt;* &lt;pointer&gt;, &lt;ty&gt; &lt;cmp&gt;, &lt;ty&gt; &lt;new&gt; [singlethread] &lt;ordering&gt; <i>; yields {ty}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4928</pre>
4929
4930<h5>Overview:</h5>
4931<p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
4932It loads a value in memory and compares it to a given value. If they are
4933equal, it stores a new value into the memory.</p>
4934
4935<h5>Arguments:</h5>
4936<p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
4937address to operate on, a value to compare to the value currently be at that
4938address, and a new value to place at that address if the compared values are
4939equal. The type of '<var>&lt;cmp&gt;</var>' must be an integer type whose
4940bit width is a power of two greater than or equal to eight and less than
4941or equal to a target-specific size limit. '<var>&lt;cmp&gt;</var>' and
4942'<var>&lt;new&gt;</var>' must have the same type, and the type of
4943'<var>&lt;pointer&gt;</var>' must be a pointer to that type. If the
4944<code>cmpxchg</code> is marked as <code>volatile</code>, then the
4945optimizer is not allowed to modify the number or order of execution
4946of this <code>cmpxchg</code> with other <a href="#volatile">volatile
4947operations</a>.</p>
4948
4949<!-- FIXME: Extend allowed types. -->
4950
4951<p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
4952<code>cmpxchg</code> synchronizes with other atomic operations.</p>
4953
4954<p>The optional "<code>singlethread</code>" argument declares that the
4955<code>cmpxchg</code> is only atomic with respect to code (usually signal
4956handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
4957cmpxchg is atomic with respect to all other code in the system.</p>
4958
4959<p>The pointer passed into cmpxchg must have alignment greater than or equal to
4960the size in memory of the operand.
4961
4962<h5>Semantics:</h5>
4963<p>The contents of memory at the location specified by the
4964'<tt>&lt;pointer&gt;</tt>' operand is read and compared to
4965'<tt>&lt;cmp&gt;</tt>'; if the read value is the equal,
4966'<tt>&lt;new&gt;</tt>' is written. The original value at the location
4967is returned.
4968
4969<p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
4970purpose of identifying <a href="#release_sequence">release sequences</a>. A
4971failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
4972parameter determined by dropping any <code>release</code> part of the
4973<code>cmpxchg</code>'s ordering.</p>
4974
4975<!--
4976FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
4977optimization work on ARM.)
4978
4979FIXME: Is a weaker ordering constraint on failure helpful in practice?
4980-->
4981
4982<h5>Example:</h5>
4983<pre>
4984entry:
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4985 %orig = atomic <a href="#i_load">load</a> i32* %ptr unordered <i>; yields {i32}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4986 <a href="#i_br">br</a> label %loop
4987
4988loop:
4989 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
4990 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4991 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]4992 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
4993 <a href="#i_br">br</a> i1 %success, label %done, label %loop
4994
4995done:
4996 ...
4997</pre>
4998
4999</div>
5000
5001<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5002<h4>
5003<a name="i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a>
5004</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5005
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5006<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5007
5008<h5>Syntax:</h5>
5009<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]5010 atomicrmw [volatile] &lt;operation&gt; &lt;ty&gt;* &lt;pointer&gt;, &lt;ty&gt; &lt;value&gt; [singlethread] &lt;ordering&gt; <i>; yields {ty}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5011</pre>
5012
5013<h5>Overview:</h5>
5014<p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
5015
5016<h5>Arguments:</h5>
5017<p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
5018operation to apply, an address whose value to modify, an argument to the
5019operation. The operation must be one of the following keywords:</p>
5020<ul>
5021 <li>xchg</li>
5022 <li>add</li>
5023 <li>sub</li>
5024 <li>and</li>
5025 <li>nand</li>
5026 <li>or</li>
5027 <li>xor</li>
5028 <li>max</li>
5029 <li>min</li>
5030 <li>umax</li>
5031 <li>umin</li>
5032</ul>
5033
5034<p>The type of '<var>&lt;value&gt;</var>' must be an integer type whose
5035bit width is a power of two greater than or equal to eight and less than
5036or equal to a target-specific size limit. The type of the
5037'<code>&lt;pointer&gt;</code>' operand must be a pointer to that type.
5038If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
5039optimizer is not allowed to modify the number or order of execution of this
5040<code>atomicrmw</code> with other <a href="#volatile">volatile
5041 operations</a>.</p>
5042
5043<!-- FIXME: Extend allowed types. -->
5044
5045<h5>Semantics:</h5>
5046<p>The contents of memory at the location specified by the
5047'<tt>&lt;pointer&gt;</tt>' operand are atomically read, modified, and written
5048back. The original value at the location is returned. The modification is
5049specified by the <var>operation</var> argument:</p>
5050
5051<ul>
5052 <li>xchg: <code>*ptr = val</code></li>
5053 <li>add: <code>*ptr = *ptr + val</code></li>
5054 <li>sub: <code>*ptr = *ptr - val</code></li>
5055 <li>and: <code>*ptr = *ptr &amp; val</code></li>
5056 <li>nand: <code>*ptr = ~(*ptr &amp; val)</code></li>
5057 <li>or: <code>*ptr = *ptr | val</code></li>
5058 <li>xor: <code>*ptr = *ptr ^ val</code></li>
5059 <li>max: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using a signed comparison)</li>
5060 <li>min: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using a signed comparison)</li>
5061 <li>umax: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5062 <li>umin: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5063</ul>
5064
5065<h5>Example:</h5>
5066<pre>
5067 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
5068</pre>
5069
5070</div>
5071
5072<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5073<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5074 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5075</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5076
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5077<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5078
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5079<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5080<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5081 &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]5082 &lt;result&gt; = getelementptr inbounds &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5083 &lt;result&gt; = getelementptr &lt;ptr vector&gt; ptrval, &lt;vector index type&gt; idx
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5084</pre>
5085
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5086<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5087<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5088 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
5089 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5090
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5091<h5>Arguments:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5092<p>The first argument is always a pointer or a vector of pointers,
5093 and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]5094 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5095 elements of the aggregate object are indexed. The interpretation of each
5096 index is dependent on the type being indexed into. The first index always
5097 indexes the pointer value given as the first argument, the second index
5098 indexes a value of the type pointed to (not necessarily the value directly
5099 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5100 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]5101 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5102 can never be pointers, since that would require loading the pointer before
5103 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5104
5105<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]5106 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5107 integer <b>constants</b> are allowed. When indexing into an array, pointer
5108 or vector, integers of any width are allowed, and they are not required to be
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5109 constant. These integers are treated as signed values where relevant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5110
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5111<p>For example, let's consider a C code fragment and how it gets compiled to
5112 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5113
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5114<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5115struct RT {
5116 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5117 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5118 char C;
5119};
5120struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5121 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5122 double Y;
5123 struct RT Z;
5124};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5125
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5126int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5127 return &amp;s[1].Z.B[5][13];
5128}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5129</pre>
5130
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5131<p>The LLVM code generated by Clang is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5132
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5133<pre class="doc_code">
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5134%struct.RT = <a href="#namedtypes">type</a> { i8, [10 x [20 x i32]], i8 }
5135%struct.ST = <a href="#namedtypes">type</a> { i32, double, %struct.RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5136
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5137define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5138entry:
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5139 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5140 ret i32* %arrayidx
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5141}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5142</pre>
5143
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5144<h5>Semantics:</h5>
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5145<p>In the example above, the first index is indexing into the
5146 '<tt>%struct.ST*</tt>' type, which is a pointer, yielding a
5147 '<tt>%struct.ST</tt>' = '<tt>{ i32, double, %struct.RT }</tt>' type, a
5148 structure. The second index indexes into the third element of the structure,
5149 yielding a '<tt>%struct.RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]], i8 }</tt>'
5150 type, another structure. The third index indexes into the second element of
5151 the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an array. The
5152 two dimensions of the array are subscripted into, yielding an '<tt>i32</tt>'
5153 type. The '<tt>getelementptr</tt>' instruction returns a pointer to this
5154 element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5155
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5156<p>Note that it is perfectly legal to index partially through a structure,
5157 returning a pointer to an inner element. Because of this, the LLVM code for
5158 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5159
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5160<pre class="doc_code">
5161define i32* @foo(%struct.ST* %s) {
5162 %t1 = getelementptr %struct.ST* %s, i32 1 <i>; yields %struct.ST*:%t1</i>
5163 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 <i>; yields %struct.RT*:%t2</i>
5164 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
5165 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
5166 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
5167 ret i32* %t5
5168}
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5169</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]5170
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5171<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]5172 <tt>getelementptr</tt> is a <a href="#poisonvalues">poison value</a> if the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]5173 base pointer is not an <i>in bounds</i> address of an allocated object,
5174 or if any of the addresses that would be formed by successive addition of
5175 the offsets implied by the indices to the base address with infinitely
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5176 precise signed arithmetic are not an <i>in bounds</i> address of that
5177 allocated object. The <i>in bounds</i> addresses for an allocated object
5178 are all the addresses that point into the object, plus the address one
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5179 byte past the end.
5180 In cases where the base is a vector of pointers the <tt>inbounds</tt> keyword
5181 applies to each of the computations element-wise. </p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5182
5183<p>If the <tt>inbounds</tt> keyword is not present, the offsets are added to
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5184 the base address with silently-wrapping two's complement arithmetic. If the
5185 offsets have a different width from the pointer, they are sign-extended or
5186 truncated to the width of the pointer. The result value of the
5187 <tt>getelementptr</tt> may be outside the object pointed to by the base
5188 pointer. The result value may not necessarily be used to access memory
5189 though, even if it happens to point into allocated storage. See the
5190 <a href="#pointeraliasing">Pointer Aliasing Rules</a> section for more
5191 information.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5192
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5193<p>The getelementptr instruction is often confusing. For some more insight into
5194 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]5195
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5196<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5197<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5198 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5199 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5200 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5201 %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]5202 <i>; yields i8*:eptr</i>
5203 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +0000[diff] [blame]5204 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +0000[diff] [blame]5205 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5206</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5207
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5208<p>In cases where the pointer argument is a vector of pointers, only a
5209 single index may be used, and the number of vector elements has to be
5210 the same. For example: </p>
5211<pre class="doc_code">
5212 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5213</pre>
5214
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5215</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5216
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5217</div>
5218
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5219<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5220<h3>
5221 <a name="convertops">Conversion Operations</a>
5222</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5223
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5224<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5225
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5226<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5227 which all take a single operand and a type. They perform various bit
5228 conversions on the operand.</p>
5229
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5230<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5231<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5232 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5233</h4>
5234
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5235<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5236
5237<h5>Syntax:</h5>
5238<pre>
5239 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5240</pre>
5241
5242<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5243<p>The '<tt>trunc</tt>' instruction truncates its operand to the
5244 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5245
5246<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5247<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5248 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5249 of the same number of integers.
5250 The bit size of the <tt>value</tt> must be larger than
5251 the bit size of the destination type, <tt>ty2</tt>.
5252 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5253
5254<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5255<p>The '<tt>trunc</tt>' instruction truncates the high order bits
5256 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5257 source size must be larger than the destination size, <tt>trunc</tt> cannot
5258 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5259
5260<h5>Example:</h5>
5261<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5262 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5263 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5264 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5265 %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]5266</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5267
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5268</div>
5269
5270<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5271<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5272 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5273</h4>
5274
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5275<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5276
5277<h5>Syntax:</h5>
5278<pre>
5279 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5280</pre>
5281
5282<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5283<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5284 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5285
5286
5287<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5288<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5289 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5290 of the same number of integers.
5291 The bit size of the <tt>value</tt> must be smaller than
5292 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5293 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5294
5295<h5>Semantics:</h5>
5296<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]5297 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5298
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]5299<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]5300
5301<h5>Example:</h5>
5302<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5303 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5304 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5305 %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]5306</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5307
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5308</div>
5309
5310<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5311<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5312 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5313</h4>
5314
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5315<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5316
5317<h5>Syntax:</h5>
5318<pre>
5319 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5320</pre>
5321
5322<h5>Overview:</h5>
5323<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5324
5325<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5326<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5327 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5328 of the same number of integers.
5329 The bit size of the <tt>value</tt> must be smaller than
5330 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5331 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5332
5333<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5334<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5335 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5336 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5337
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5338<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]5339
5340<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5341<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5342 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5343 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5344 %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]5345</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5346
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5347</div>
5348
5349<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5350<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5351 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5352</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5353
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5354<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5355
5356<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5357<pre>
5358 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5359</pre>
5360
5361<h5>Overview:</h5>
5362<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5363 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5364
5365<h5>Arguments:</h5>
5366<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5367 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5368 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]5369 <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]5370 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5371
5372<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5373<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5374 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5375 <a href="#t_floating">floating point</a> type. If the value cannot fit
5376 within the destination type, <tt>ty2</tt>, then the results are
5377 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5378
5379<h5>Example:</h5>
5380<pre>
5381 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5382 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5383</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5384
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5385</div>
5386
5387<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5388<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5389 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5390</h4>
5391
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5392<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5393
5394<h5>Syntax:</h5>
5395<pre>
5396 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5397</pre>
5398
5399<h5>Overview:</h5>
5400<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]5401 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5402
5403<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5404<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5405 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5406 a <a href="#t_floating">floating point</a> type to cast it to. The source
5407 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5408
5409<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5410<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5411 <a href="#t_floating">floating point</a> type to a larger
5412 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5413 used to make a <i>no-op cast</i> because it always changes bits. Use
5414 <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]5415
5416<h5>Example:</h5>
5417<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +0000[diff] [blame]5418 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5419 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5420</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5421
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5422</div>
5423
5424<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5425<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]5426 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5427</h4>
5428
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5429<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5430
5431<h5>Syntax:</h5>
5432<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5433 &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]5434</pre>
5435
5436<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5437<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5438 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5439
5440<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5441<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5442 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5443 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5444 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5445 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5446
5447<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5448<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5449 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5450 towards zero) unsigned integer value. If the value cannot fit
5451 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5452
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5453<h5>Example:</h5>
5454<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5455 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5456 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5457 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5458</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5459
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5460</div>
5461
5462<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5463<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5464 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5465</h4>
5466
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5467<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5468
5469<h5>Syntax:</h5>
5470<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5471 &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]5472</pre>
5473
5474<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5475<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5476 <a href="#t_floating">floating point</a> <tt>value</tt> to
5477 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5478
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5479<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5480<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5481 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5482 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5483 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5484 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5485
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5486<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5487<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5488 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5489 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5490 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5491
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5492<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5493<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5494 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5495 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5496 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5497</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5498
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5499</div>
5500
5501<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5502<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5503 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5504</h4>
5505
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5506<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5507
5508<h5>Syntax:</h5>
5509<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5510 &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]5511</pre>
5512
5513<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5514<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5515 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5516
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5517<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5518<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]5519 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5520 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5521 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5522 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5523
5524<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5525<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5526 integer quantity and converts it to the corresponding floating point
5527 value. If the value cannot fit in the floating point value, the results are
5528 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5529
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5530<h5>Example:</h5>
5531<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5532 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5533 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5534</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5535
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5536</div>
5537
5538<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5539<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5540 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5541</h4>
5542
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5543<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5544
5545<h5>Syntax:</h5>
5546<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5547 &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]5548</pre>
5549
5550<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5551<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5552 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5553
5554<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5555<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]5556 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5557 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5558 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5559 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5560
5561<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5562<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5563 quantity and converts it to the corresponding floating point value. If the
5564 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5565
5566<h5>Example:</h5>
5567<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5568 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5569 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5570</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5571
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5572</div>
5573
5574<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5575<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5576 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5577</h4>
5578
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5579<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5580
5581<h5>Syntax:</h5>
5582<pre>
5583 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5584</pre>
5585
5586<h5>Overview:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5587<p>The '<tt>ptrtoint</tt>' instruction converts the pointer or a vector of
5588 pointers <tt>value</tt> to
5589 the integer (or vector of integers) type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5590
5591<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5592<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5593 must be a a value of type <a href="#t_pointer">pointer</a> or a vector of
5594 pointers, and a type to cast it to
5595 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> or a vector
5596 of integers type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5597
5598<h5>Semantics:</h5>
5599<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5600 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5601 truncating or zero extending that value to the size of the integer type. If
5602 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5603 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5604 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5605 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5606
5607<h5>Example:</h5>
5608<pre>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5609 %X = ptrtoint i32* %P to i8 <i>; yields truncation on 32-bit architecture</i>
5610 %Y = ptrtoint i32* %P to i64 <i>; yields zero extension on 32-bit architecture</i>
5611 %Z = ptrtoint &lt;4 x i32*&gt; %P to &lt;4 x i64&gt;<i>; yields vector zero extension for a vector of addresses on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5612</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5613
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5614</div>
5615
5616<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5617<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5618 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5619</h4>
5620
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5621<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5622
5623<h5>Syntax:</h5>
5624<pre>
5625 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5626</pre>
5627
5628<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5629<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5630 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5631
5632<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]5633<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5634 value to cast, and a type to cast it to, which must be a
5635 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5636
5637<h5>Semantics:</h5>
5638<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5639 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5640 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5641 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5642 than the size of a pointer then a zero extension is done. If they are the
5643 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5644
5645<h5>Example:</h5>
5646<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5647 %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]5648 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5649 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5650 %Z = inttoptr &lt;4 x i32&gt; %G to &lt;4 x i8*&gt;<i>; yields truncation of vector G to four pointers</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5651</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5652
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5653</div>
5654
5655<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5656<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5657 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5658</h4>
5659
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5660<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5661
5662<h5>Syntax:</h5>
5663<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5664 &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]5665</pre>
5666
5667<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5668<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5669 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5670
5671<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5672<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5673 non-aggregate first class value, and a type to cast it to, which must also be
5674 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5675 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5676 identical. If the source type is a pointer, the destination type must also be
5677 a pointer. This instruction supports bitwise conversion of vectors to
5678 integers and to vectors of other types (as long as they have the same
5679 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5680
5681<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5682<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5683 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5684 this conversion. The conversion is done as if the <tt>value</tt> had been
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5685 stored to memory and read back as type <tt>ty2</tt>.
5686 Pointer (or vector of pointers) types may only be converted to other pointer
5687 (or vector of pointers) types with this instruction. To convert
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5688 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5689 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5690
5691<h5>Example:</h5>
5692<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5693 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5694 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5695 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
5696 %Z = bitcast &lt;2 x i32*&gt; %V to &lt;2 x i64*&gt; <i>; yields &lt;2 x i64*&gt;</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5697</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5698
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5699</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5700
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5701</div>
5702
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5703<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5704<h3>
5705 <a name="otherops">Other Operations</a>
5706</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5707
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5708<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5709
5710<p>The instructions in this category are the "miscellaneous" instructions, which
5711 defy better classification.</p>
5712
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5713<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5714<h4>
5715 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5716</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5717
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5718<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5719
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5720<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5721<pre>
5722 &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]5723</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5724
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5725<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5726<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5727 boolean values based on comparison of its two integer, integer vector,
5728 pointer, or pointer vector operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5729
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5730<h5>Arguments:</h5>
5731<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5732 the condition code indicating the kind of comparison to perform. It is not a
5733 value, just a keyword. The possible condition code are:</p>
5734
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5735<ol>
5736 <li><tt>eq</tt>: equal</li>
5737 <li><tt>ne</tt>: not equal </li>
5738 <li><tt>ugt</tt>: unsigned greater than</li>
5739 <li><tt>uge</tt>: unsigned greater or equal</li>
5740 <li><tt>ult</tt>: unsigned less than</li>
5741 <li><tt>ule</tt>: unsigned less or equal</li>
5742 <li><tt>sgt</tt>: signed greater than</li>
5743 <li><tt>sge</tt>: signed greater or equal</li>
5744 <li><tt>slt</tt>: signed less than</li>
5745 <li><tt>sle</tt>: signed less or equal</li>
5746</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5747
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]5748<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5749 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
5750 typed. They must also be identical types.</p>
5751
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5752<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5753<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
5754 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5755 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]5756 result, as follows:</p>
5757
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5758<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5759 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5760 <tt>false</tt> otherwise. No sign interpretation is necessary or
5761 performed.</li>
5762
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5763 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5764 <tt>false</tt> otherwise. No sign interpretation is necessary or
5765 performed.</li>
5766
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5767 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5768 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5769
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5770 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5771 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5772 to <tt>op2</tt>.</li>
5773
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5774 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5775 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5776
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5777 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5778 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5779
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5780 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5781 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5782
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5783 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5784 <tt>true</tt> if <tt>op1</tt> is greater than or equal
5785 to <tt>op2</tt>.</li>
5786
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5787 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5788 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
5789
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5790 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5791 <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]5792</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5793
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5794<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5795 values are compared as if they were integers.</p>
5796
5797<p>If the operands are integer vectors, then they are compared element by
5798 element. The result is an <tt>i1</tt> vector with the same number of elements
5799 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5800
5801<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5802<pre>
5803 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5804 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
5805 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
5806 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
5807 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
5808 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5809</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5810
5811<p>Note that the code generator does not yet support vector types with
5812 the <tt>icmp</tt> instruction.</p>
5813
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5814</div>
5815
5816<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5817<h4>
5818 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
5819</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5820
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5821<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5822
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5823<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5824<pre>
5825 &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]5826</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5827
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5828<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5829<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
5830 values based on comparison of its operands.</p>
5831
5832<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]5833(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5834
5835<p>If the operands are floating point vectors, then the result type is a vector
5836 of boolean with the same number of elements as the operands being
5837 compared.</p>
5838
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5839<h5>Arguments:</h5>
5840<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5841 the condition code indicating the kind of comparison to perform. It is not a
5842 value, just a keyword. The possible condition code are:</p>
5843
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5844<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5845 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5846 <li><tt>oeq</tt>: ordered and equal</li>
5847 <li><tt>ogt</tt>: ordered and greater than </li>
5848 <li><tt>oge</tt>: ordered and greater than or equal</li>
5849 <li><tt>olt</tt>: ordered and less than </li>
5850 <li><tt>ole</tt>: ordered and less than or equal</li>
5851 <li><tt>one</tt>: ordered and not equal</li>
5852 <li><tt>ord</tt>: ordered (no nans)</li>
5853 <li><tt>ueq</tt>: unordered or equal</li>
5854 <li><tt>ugt</tt>: unordered or greater than </li>
5855 <li><tt>uge</tt>: unordered or greater than or equal</li>
5856 <li><tt>ult</tt>: unordered or less than </li>
5857 <li><tt>ule</tt>: unordered or less than or equal</li>
5858 <li><tt>une</tt>: unordered or not equal</li>
5859 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5860 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5861</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5862
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5863<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5864 <i>unordered</i> means that either operand may be a QNAN.</p>
5865
5866<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
5867 a <a href="#t_floating">floating point</a> type or
5868 a <a href="#t_vector">vector</a> of floating point type. They must have
5869 identical types.</p>
5870
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5871<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]5872<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5873 according to the condition code given as <tt>cond</tt>. If the operands are
5874 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]5875 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5876 follows:</p>
5877
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5878<ol>
5879 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5880
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5881 <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]5882 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5883
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5884 <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]5885 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5886
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5887 <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]5888 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5889
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5890 <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]5891 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5892
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5893 <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]5894 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5895
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5896 <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]5897 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5898
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5899 <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]5900
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5901 <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]5902 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
5903
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5904 <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]5905 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
5906
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5907 <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]5908 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
5909
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5910 <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]5911 <tt>op1</tt> is less than <tt>op2</tt>.</li>
5912
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5913 <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]5914 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
5915
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5916 <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]5917 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
5918
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]5919 <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]5920
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5921 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
5922</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5923
5924<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5925<pre>
5926 &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]5927 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
5928 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
5929 &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]5930</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]5931
5932<p>Note that the code generator does not yet support vector types with
5933 the <tt>fcmp</tt> instruction.</p>
5934
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5935</div>
5936
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5937<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5938<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5939 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5940</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5941
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5942<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5943
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5944<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5945<pre>
5946 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
5947</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5948
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5949<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5950<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
5951 SSA graph representing the function.</p>
5952
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5953<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5954<p>The type of the incoming values is specified with the first type field. After
5955 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
5956 one pair for each predecessor basic block of the current block. Only values
5957 of <a href="#t_firstclass">first class</a> type may be used as the value
5958 arguments to the PHI node. Only labels may be used as the label
5959 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5960
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5961<p>There must be no non-phi instructions between the start of a basic block and
5962 the PHI instructions: i.e. PHI instructions must be first in a basic
5963 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5964
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5965<p>For the purposes of the SSA form, the use of each incoming value is deemed to
5966 occur on the edge from the corresponding predecessor block to the current
5967 block (but after any definition of an '<tt>invoke</tt>' instruction's return
5968 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +0000[diff] [blame]5969
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5970<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5971<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5972 specified by the pair corresponding to the predecessor basic block that
5973 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5974
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5975<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]5976<pre>
5977Loop: ; Infinite loop that counts from 0 on up...
5978 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
5979 %nextindvar = add i32 %indvar, 1
5980 br label %Loop
5981</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5982
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5983</div>
5984
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5985<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5986<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5987 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5988</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5989
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5990<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5991
5992<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5993<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]5994 &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>
5995
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5996 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5997</pre>
5998
5999<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6000<p>The '<tt>select</tt>' instruction is used to choose one value based on a
6001 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6002
6003
6004<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6005<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
6006 values indicating the condition, and two values of the
6007 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
6008 vectors and the condition is a scalar, then entire vectors are selected, not
6009 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6010
6011<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6012<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
6013 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6014
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6015<p>If the condition is a vector of i1, then the value arguments must be vectors
6016 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6017
6018<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6019<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]6020 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6021</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6022
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6023</div>
6024
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]6025<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6026<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6027 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6028</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6029
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6030<div>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6031
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6032<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6033<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]6034 &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]6035</pre>
6036
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6037<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6038<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6039
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6040<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6041<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6042
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6043<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6044 <li>The optional "tail" marker indicates that the callee function does not
6045 access any allocas or varargs in the caller. Note that calls may be
6046 marked "tail" even if they do not occur before
6047 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
6048 present, the function call is eligible for tail call optimization,
6049 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +0000[diff] [blame]6050 optimized into a jump</a>. The code generator may optimize calls marked
6051 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
6052 sibling call optimization</a> when the caller and callee have
6053 matching signatures, or 2) forced tail call optimization when the
6054 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6055 <ul>
6056 <li>Caller and callee both have the calling
6057 convention <tt>fastcc</tt>.</li>
6058 <li>The call is in tail position (ret immediately follows call and ret
6059 uses value of call or is void).</li>
6060 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +0000[diff] [blame]6061 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6062 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
6063 constraints are met.</a></li>
6064 </ul>
6065 </li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6066
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6067 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
6068 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6069 defaults to using C calling conventions. The calling convention of the
6070 call must match the calling convention of the target function, or else the
6071 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6072
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6073 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
6074 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
6075 '<tt>inreg</tt>' attributes are valid here.</li>
6076
6077 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
6078 type of the return value. Functions that return no value are marked
6079 <tt><a href="#t_void">void</a></tt>.</li>
6080
6081 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
6082 being invoked. The argument types must match the types implied by this
6083 signature. This type can be omitted if the function is not varargs and if
6084 the function type does not return a pointer to a function.</li>
6085
6086 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
6087 be invoked. In most cases, this is a direct function invocation, but
6088 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
6089 to function value.</li>
6090
6091 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]6092 signature argument types and parameter attributes. All arguments must be
6093 of <a href="#t_firstclass">first class</a> type. If the function
6094 signature indicates the function accepts a variable number of arguments,
6095 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6096
6097 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
6098 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
6099 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6100</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6101
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6102<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6103<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
6104 a specified function, with its incoming arguments bound to the specified
6105 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
6106 function, control flow continues with the instruction after the function
6107 call, and the return value of the function is bound to the result
6108 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6109
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6110<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6111<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]6112 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6113 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +0000[diff] [blame]6114 %X = tail call i32 @foo() <i>; yields i32</i>
6115 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
6116 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]6117
6118 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6119 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]6120 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
6121 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +0000[diff] [blame]6122 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +0000[diff] [blame]6123 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6124</pre>
6125
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6126<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +0000[diff] [blame]6127standard C99 library as being the C99 library functions, and may perform
6128optimizations or generate code for them under that assumption. This is
6129something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6130freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6131
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6132</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6133
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6134<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6135<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]6136 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6137</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6138
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6139<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6140
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6141<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6142<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6143 &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]6144</pre>
6145
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6146<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6147<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6148 the "variable argument" area of a function call. It is used to implement the
6149 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6150
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6151<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6152<p>This instruction takes a <tt>va_list*</tt> value and the type of the
6153 argument. It returns a value of the specified argument type and increments
6154 the <tt>va_list</tt> to point to the next argument. The actual type
6155 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6156
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6157<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6158<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
6159 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
6160 to the next argument. For more information, see the variable argument
6161 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6162
6163<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]6164 take a variable number of arguments, for example, the <tt>vfprintf</tt>
6165 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6166
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6167<p><tt>va_arg</tt> is an LLVM instruction instead of
6168 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
6169 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6170
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6171<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6172<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
6173
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6174<p>Note that the code generator does not yet fully support va_arg on many
6175 targets. Also, it does not currently support va_arg with aggregate types on
6176 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]6177
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6178</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6179
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6180<!-- _______________________________________________________________________ -->
6181<h4>
6182 <a name="i_landingpad">'<tt>landingpad</tt>' Instruction</a>
6183</h4>
6184
6185<div>
6186
6187<h5>Syntax:</h5>
6188<pre>
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6189 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; &lt;clause&gt;+
6190 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; cleanup &lt;clause&gt;*
Bill Wendlingbf13ee12011-08-08 08:06:05 +0000[diff] [blame]6191
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6192 &lt;clause&gt; := catch &lt;type&gt; &lt;value&gt;
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6193 &lt;clause&gt; := filter &lt;array constant type&gt; &lt;array constant&gt;
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6194</pre>
6195
6196<h5>Overview:</h5>
6197<p>The '<tt>landingpad</tt>' instruction is used by
6198 <a href="ExceptionHandling.html#overview">LLVM's exception handling
6199 system</a> to specify that a basic block is a landing pad &mdash; one where
6200 the exception lands, and corresponds to the code found in the
6201 <i><tt>catch</tt></i> portion of a <i><tt>try/catch</tt></i> sequence. It
6202 defines values supplied by the personality function (<tt>pers_fn</tt>) upon
6203 re-entry to the function. The <tt>resultval</tt> has the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6204 type <tt>resultty</tt>.</p>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6205
6206<h5>Arguments:</h5>
6207<p>This instruction takes a <tt>pers_fn</tt> value. This is the personality
6208 function associated with the unwinding mechanism. The optional
6209 <tt>cleanup</tt> flag indicates that the landing pad block is a cleanup.</p>
6210
6211<p>A <tt>clause</tt> begins with the clause type &mdash; <tt>catch</tt>
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6212 or <tt>filter</tt> &mdash; and contains the global variable representing the
6213 "type" that may be caught or filtered respectively. Unlike the
6214 <tt>catch</tt> clause, the <tt>filter</tt> clause takes an array constant as
6215 its argument. Use "<tt>[0 x i8**] undef</tt>" for a filter which cannot
6216 throw. The '<tt>landingpad</tt>' instruction must contain <em>at least</em>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6217 one <tt>clause</tt> or the <tt>cleanup</tt> flag.</p>
6218
6219<h5>Semantics:</h5>
6220<p>The '<tt>landingpad</tt>' instruction defines the values which are set by the
6221 personality function (<tt>pers_fn</tt>) upon re-entry to the function, and
6222 therefore the "result type" of the <tt>landingpad</tt> instruction. As with
6223 calling conventions, how the personality function results are represented in
6224 LLVM IR is target specific.</p>
6225
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6226<p>The clauses are applied in order from top to bottom. If two
6227 <tt>landingpad</tt> instructions are merged together through inlining, the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6228 clauses from the calling function are appended to the list of clauses.
6229 When the call stack is being unwound due to an exception being thrown, the
6230 exception is compared against each <tt>clause</tt> in turn. If it doesn't
6231 match any of the clauses, and the <tt>cleanup</tt> flag is not set, then
6232 unwinding continues further up the call stack.</p>
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6233
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6234<p>The <tt>landingpad</tt> instruction has several restrictions:</p>
6235
6236<ul>
6237 <li>A landing pad block is a basic block which is the unwind destination of an
6238 '<tt>invoke</tt>' instruction.</li>
6239 <li>A landing pad block must have a '<tt>landingpad</tt>' instruction as its
6240 first non-PHI instruction.</li>
6241 <li>There can be only one '<tt>landingpad</tt>' instruction within the landing
6242 pad block.</li>
6243 <li>A basic block that is not a landing pad block may not include a
6244 '<tt>landingpad</tt>' instruction.</li>
6245 <li>All '<tt>landingpad</tt>' instructions in a function must have the same
6246 personality function.</li>
6247</ul>
6248
6249<h5>Example:</h5>
6250<pre>
6251 ;; A landing pad which can catch an integer.
6252 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6253 catch i8** @_ZTIi
6254 ;; A landing pad that is a cleanup.
6255 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6256 cleanup
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6257 ;; A landing pad which can catch an integer and can only throw a double.
6258 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6259 catch i8** @_ZTIi
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6260 filter [1 x i8**] [@_ZTId]
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6261</pre>
6262
6263</div>
6264
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6265</div>
6266
6267</div>
6268
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6269<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6270<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6271<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6272
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6273<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6274
6275<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6276 well known names and semantics and are required to follow certain
6277 restrictions. Overall, these intrinsics represent an extension mechanism for
6278 the LLVM language that does not require changing all of the transformations
6279 in LLVM when adding to the language (or the bitcode reader/writer, the
6280 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6281
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]6282<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6283 prefix is reserved in LLVM for intrinsic names; thus, function names may not
6284 begin with this prefix. Intrinsic functions must always be external
6285 functions: you cannot define the body of intrinsic functions. Intrinsic
6286 functions may only be used in call or invoke instructions: it is illegal to
6287 take the address of an intrinsic function. Additionally, because intrinsic
6288 functions are part of the LLVM language, it is required if any are added that
6289 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6290
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6291<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6292 family of functions that perform the same operation but on different data
6293 types. Because LLVM can represent over 8 million different integer types,
6294 overloading is used commonly to allow an intrinsic function to operate on any
6295 integer type. One or more of the argument types or the result type can be
6296 overloaded to accept any integer type. Argument types may also be defined as
6297 exactly matching a previous argument's type or the result type. This allows
6298 an intrinsic function which accepts multiple arguments, but needs all of them
6299 to be of the same type, to only be overloaded with respect to a single
6300 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6301
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6302<p>Overloaded intrinsics will have the names of its overloaded argument types
6303 encoded into its function name, each preceded by a period. Only those types
6304 which are overloaded result in a name suffix. Arguments whose type is matched
6305 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6306 can take an integer of any width and returns an integer of exactly the same
6307 integer width. This leads to a family of functions such as
6308 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6309 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6310 suffix is required. Because the argument's type is matched against the return
6311 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6312
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6313<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6314 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6315
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6316<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6317<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6318 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6319</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6320
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6321<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6322
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6323<p>Variable argument support is defined in LLVM with
6324 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6325 intrinsic functions. These functions are related to the similarly named
6326 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6327
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6328<p>All of these functions operate on arguments that use a target-specific value
6329 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6330 not define what this type is, so all transformations should be prepared to
6331 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6332
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]6333<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]6334 instruction and the variable argument handling intrinsic functions are
6335 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6336
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]6337<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6338define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6339 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6340 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6341 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6342 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6343
6344 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6345 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6346
6347 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6348 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6349 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6350 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6351 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6352
6353 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6354 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]6355 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6356}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6357
6358declare void @llvm.va_start(i8*)
6359declare void @llvm.va_copy(i8*, i8*)
6360declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6361</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6362
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6363<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6364<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6365 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6366</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6367
6368
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6369<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6370
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6371<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6372<pre>
6373 declare void %llvm.va_start(i8* &lt;arglist&gt;)
6374</pre>
6375
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6376<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6377<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
6378 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]6379
6380<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6381<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]6382
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6383<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6384<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]6385 macro available in C. In a target-dependent way, it initializes
6386 the <tt>va_list</tt> element to which the argument points, so that the next
6387 call to <tt>va_arg</tt> will produce the first variable argument passed to
6388 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6389 need to know the last argument of the function as the compiler can figure
6390 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6391
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6392</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6393
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6394<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6395<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6396 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6397</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6398
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6399<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6400
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6401<h5>Syntax:</h5>
6402<pre>
6403 declare void @llvm.va_end(i8* &lt;arglist&gt;)
6404</pre>
6405
6406<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6407<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6408 which has been initialized previously
6409 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6410 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6411
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6412<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6413<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]6414
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6415<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6416<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]6417 macro available in C. In a target-dependent way, it destroys
6418 the <tt>va_list</tt> element to which the argument points. Calls
6419 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6420 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6421 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6422
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6423</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6424
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6425<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6426<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6427 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6428</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6429
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6430<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6431
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6432<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6433<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6434 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6435</pre>
6436
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6437<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6438<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6439 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6440
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6441<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6442<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]6443 The second argument is a pointer to a <tt>va_list</tt> element to copy
6444 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6445
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6446<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6447<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]6448 macro available in C. In a target-dependent way, it copies the
6449 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6450 element. This intrinsic is necessary because
6451 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6452 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6453
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6454</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6455
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6456</div>
6457
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6458<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6459<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6460 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6461</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6462
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6463<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6464
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6465<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]6466Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6467intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6468roots on the stack</a>, as well as garbage collector implementations that
6469require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6470barriers. Front-ends for type-safe garbage collected languages should generate
6471these intrinsics to make use of the LLVM garbage collectors. For more details,
6472see <a href="GarbageCollection.html">Accurate Garbage Collection with
6473LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6474
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6475<p>The garbage collection intrinsics only operate on objects in the generic
6476 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6477
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6478<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6479<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6480 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6481</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6482
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6483<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6484
6485<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6486<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6487 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6488</pre>
6489
6490<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]6491<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]6492 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6493
6494<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6495<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6496 root pointer. The second pointer (which must be either a constant or a
6497 global value address) contains the meta-data to be associated with the
6498 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6499
6500<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]6501<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]6502 location. At compile-time, the code generator generates information to allow
6503 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6504 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6505 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6506
6507</div>
6508
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6509<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6510<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6511 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6512</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6513
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6514<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6515
6516<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6517<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6518 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6519</pre>
6520
6521<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6522<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6523 locations, allowing garbage collector implementations that require read
6524 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6525
6526<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6527<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]6528 allocated from the garbage collector. The first object is a pointer to the
6529 start of the referenced object, if needed by the language runtime (otherwise
6530 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6531
6532<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6533<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6534 instruction, but may be replaced with substantially more complex code by the
6535 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6536 may only be used in a function which <a href="#gc">specifies a GC
6537 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6538
6539</div>
6540
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6541<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6542<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6543 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6544</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6545
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6546<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6547
6548<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6549<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6550 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6551</pre>
6552
6553<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6554<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6555 locations, allowing garbage collector implementations that require write
6556 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6557
6558<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6559<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]6560 object to store it to, and the third is the address of the field of Obj to
6561 store to. If the runtime does not require a pointer to the object, Obj may
6562 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6563
6564<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6565<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6566 instruction, but may be replaced with substantially more complex code by the
6567 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6568 may only be used in a function which <a href="#gc">specifies a GC
6569 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6570
6571</div>
6572
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6573</div>
6574
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6575<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6576<h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6577 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6578</h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6579
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6580<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6581
6582<p>These intrinsics are provided by LLVM to expose special features that may
6583 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6584
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6585<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6586<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6587 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6588</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6589
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6590<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6591
6592<h5>Syntax:</h5>
6593<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6594 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6595</pre>
6596
6597<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6598<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6599 target-specific value indicating the return address of the current function
6600 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6601
6602<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6603<p>The argument to this intrinsic indicates which function to return the address
6604 for. Zero indicates the calling function, one indicates its caller, etc.
6605 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6606
6607<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6608<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6609 indicating the return address of the specified call frame, or zero if it
6610 cannot be identified. The value returned by this intrinsic is likely to be
6611 incorrect or 0 for arguments other than zero, so it should only be used for
6612 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6613
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6614<p>Note that calling this intrinsic does not prevent function inlining or other
6615 aggressive transformations, so the value returned may not be that of the
6616 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6617
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6618</div>
6619
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6620<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6621<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6622 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6623</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6624
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6625<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6626
6627<h5>Syntax:</h5>
6628<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6629 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6630</pre>
6631
6632<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6633<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6634 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6635
6636<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6637<p>The argument to this intrinsic indicates which function to return the frame
6638 pointer for. Zero indicates the calling function, one indicates its caller,
6639 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6640
6641<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6642<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6643 indicating the frame address of the specified call frame, or zero if it
6644 cannot be identified. The value returned by this intrinsic is likely to be
6645 incorrect or 0 for arguments other than zero, so it should only be used for
6646 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6647
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6648<p>Note that calling this intrinsic does not prevent function inlining or other
6649 aggressive transformations, so the value returned may not be that of the
6650 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6651
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6652</div>
6653
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6654<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6655<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6656 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6657</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6658
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6659<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6660
6661<h5>Syntax:</h5>
6662<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6663 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6664</pre>
6665
6666<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6667<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6668 of the function stack, for use
6669 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6670 useful for implementing language features like scoped automatic variable
6671 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6672
6673<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6674<p>This intrinsic returns a opaque pointer value that can be passed
6675 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6676 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6677 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6678 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6679 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6680 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6681
6682</div>
6683
6684<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6685<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6686 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6687</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6688
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6689<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6690
6691<h5>Syntax:</h5>
6692<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6693 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6694</pre>
6695
6696<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6697<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6698 the function stack to the state it was in when the
6699 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6700 executed. This is useful for implementing language features like scoped
6701 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6702
6703<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6704<p>See the description
6705 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6706
6707</div>
6708
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6709<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6710<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6711 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6712</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6713
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6714<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6715
6716<h5>Syntax:</h5>
6717<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6718 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]6719</pre>
6720
6721<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6722<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6723 insert a prefetch instruction if supported; otherwise, it is a noop.
6724 Prefetches have no effect on the behavior of the program but can change its
6725 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6726
6727<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6728<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
6729 specifier determining if the fetch should be for a read (0) or write (1),
6730 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6731 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
6732 specifies whether the prefetch is performed on the data (1) or instruction (0)
6733 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
6734 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6735
6736<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6737<p>This intrinsic does not modify the behavior of the program. In particular,
6738 prefetches cannot trap and do not produce a value. On targets that support
6739 this intrinsic, the prefetch can provide hints to the processor cache for
6740 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6741
6742</div>
6743
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6744<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6745<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6746 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6747</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6748
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6749<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6750
6751<h5>Syntax:</h5>
6752<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6753 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +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.pcmarker</tt>' intrinsic is a method to export a Program
6758 Counter (PC) in a region of code to simulators and other tools. The method
6759 is target specific, but it is expected that the marker will use exported
6760 symbols to transmit the PC of the marker. The marker makes no guarantees
6761 that it will remain with any specific instruction after optimizations. It is
6762 possible that the presence of a marker will inhibit optimizations. The
6763 intended use is to be inserted after optimizations to allow correlations of
6764 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6765
6766<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6767<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6768
6769<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6770<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6771 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]6772
6773</div>
6774
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6775<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6776<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6777 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6778</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6779
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6780<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6781
6782<h5>Syntax:</h5>
6783<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]6784 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6785</pre>
6786
6787<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6788<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
6789 counter register (or similar low latency, high accuracy clocks) on those
6790 targets that support it. On X86, it should map to RDTSC. On Alpha, it
6791 should map to RPCC. As the backing counters overflow quickly (on the order
6792 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6793
6794<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6795<p>When directly supported, reading the cycle counter should not modify any
6796 memory. Implementations are allowed to either return a application specific
6797 value or a system wide value. On backends without support, this is lowered
6798 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]6799
6800</div>
6801
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6802</div>
6803
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6804<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6805<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6806 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6807</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6808
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6809<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6810
6811<p>LLVM provides intrinsics for a few important standard C library functions.
6812 These intrinsics allow source-language front-ends to pass information about
6813 the alignment of the pointer arguments to the code generator, providing
6814 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6815
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6816<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6817<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6818 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6819</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6820
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6821<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6822
6823<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6824<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]6825 integer bit width and for different address spaces. Not all targets support
6826 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6827
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6828<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6829 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]6830 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6831 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]6832 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6833</pre>
6834
6835<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6836<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6837 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6838
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6839<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6840 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6841 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6842
6843<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6844
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6845<p>The first argument is a pointer to the destination, the second is a pointer
6846 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6847 number of bytes to copy, the fourth argument is the alignment of the
6848 source and destination locations, and the fifth is a boolean indicating a
6849 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6850
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6851<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]6852 then the caller guarantees that both the source and destination pointers are
6853 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6854
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6855<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6856 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
6857 The detailed access behavior is not very cleanly specified and it is unwise
6858 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6859
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6860<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6861
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6862<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
6863 source location to the destination location, which are not allowed to
6864 overlap. It copies "len" bytes of memory over. If the argument is known to
6865 be aligned to some boundary, this can be specified as the fourth argument,
6866 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6867
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6868</div>
6869
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6870<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6871<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6872 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6873</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6874
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6875<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6876
6877<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6878<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]6879 width and for different address space. Not all targets support all bit
6880 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6881
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6882<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6883 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]6884 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6885 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]6886 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6887</pre>
6888
6889<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6890<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
6891 source location to the destination location. It is similar to the
6892 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
6893 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6894
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6895<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6896 intrinsics do not return a value, takes extra alignment/isvolatile arguments
6897 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6898
6899<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6900
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6901<p>The first argument is a pointer to the destination, the second is a pointer
6902 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6903 number of bytes to copy, the fourth argument is the alignment of the
6904 source and destination locations, and the fifth is a boolean indicating a
6905 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6906
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6907<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]6908 then the caller guarantees that the source and destination pointers are
6909 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]6910
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6911<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6912 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
6913 The detailed access behavior is not very cleanly specified and it is unwise
6914 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6915
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6916<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6917
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6918<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
6919 source location to the destination location, which may overlap. It copies
6920 "len" bytes of memory over. If the argument is known to be aligned to some
6921 boundary, this can be specified as the fourth argument, otherwise it should
6922 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6923
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]6924</div>
6925
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6926<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6927<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6928 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6929</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6930
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6931<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6932
6933<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]6934<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]6935 width and for different address spaces. However, not all targets support all
6936 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6937
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6938<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6939 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6940 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6941 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]6942 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6943</pre>
6944
6945<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6946<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
6947 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6948
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6949<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6950 intrinsic does not return a value and takes extra alignment/volatile
6951 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6952
6953<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6954<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]6955 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6956 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]6957 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6958
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6959<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]6960 then the caller guarantees that the destination pointer is aligned to that
6961 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6962
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]6963<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
6964 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
6965 The detailed access behavior is not very cleanly specified and it is unwise
6966 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]6967
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6968<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6969<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
6970 at the destination location. If the argument is known to be aligned to some
6971 boundary, this can be specified as the fourth argument, otherwise it should
6972 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6973
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6974</div>
6975
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]6976<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6977<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6978 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6979</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6980
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6981<div>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6982
6983<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6984<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
6985 floating point or vector of floating point type. Not all targets support all
6986 types however.</p>
6987
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6988<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]6989 declare float @llvm.sqrt.f32(float %Val)
6990 declare double @llvm.sqrt.f64(double %Val)
6991 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
6992 declare fp128 @llvm.sqrt.f128(fp128 %Val)
6993 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]6994</pre>
6995
6996<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6997<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
6998 returning the same value as the libm '<tt>sqrt</tt>' functions would.
6999 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
7000 behavior for negative numbers other than -0.0 (which allows for better
7001 optimization, because there is no need to worry about errno being
7002 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]7003
7004<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7005<p>The argument and return value are floating point numbers of the same
7006 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7007
7008<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7009<p>This function returns the sqrt of the specified operand if it is a
7010 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7011
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7012</div>
7013
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7014<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7015<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7016 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7017</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7018
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7019<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7020
7021<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7022<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
7023 floating point or vector of floating point type. Not all targets support all
7024 types however.</p>
7025
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7026<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]7027 declare float @llvm.powi.f32(float %Val, i32 %power)
7028 declare double @llvm.powi.f64(double %Val, i32 %power)
7029 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7030 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7031 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7032</pre>
7033
7034<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7035<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
7036 specified (positive or negative) power. The order of evaluation of
7037 multiplications is not defined. When a vector of floating point type is
7038 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7039
7040<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7041<p>The second argument is an integer power, and the first is a value to raise to
7042 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7043
7044<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7045<p>This function returns the first value raised to the second power with an
7046 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7047
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7048</div>
7049
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7050<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7051<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7052 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7053</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7054
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7055<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7056
7057<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7058<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
7059 floating point or vector of floating point type. Not all targets support all
7060 types however.</p>
7061
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7062<pre>
7063 declare float @llvm.sin.f32(float %Val)
7064 declare double @llvm.sin.f64(double %Val)
7065 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7066 declare fp128 @llvm.sin.f128(fp128 %Val)
7067 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7068</pre>
7069
7070<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7071<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7072
7073<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7074<p>The argument and return value are floating point numbers of the same
7075 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7076
7077<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7078<p>This function returns the sine of the specified operand, returning the same
7079 values as the libm <tt>sin</tt> functions would, and handles error conditions
7080 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7081
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7082</div>
7083
7084<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7085<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7086 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7087</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7088
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7089<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7090
7091<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7092<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
7093 floating point or vector of floating point type. Not all targets support all
7094 types however.</p>
7095
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7096<pre>
7097 declare float @llvm.cos.f32(float %Val)
7098 declare double @llvm.cos.f64(double %Val)
7099 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7100 declare fp128 @llvm.cos.f128(fp128 %Val)
7101 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7102</pre>
7103
7104<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7105<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7106
7107<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7108<p>The argument and return value are floating point numbers of the same
7109 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7110
7111<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7112<p>This function returns the cosine of the specified operand, returning the same
7113 values as the libm <tt>cos</tt> functions would, and handles error conditions
7114 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7115
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7116</div>
7117
7118<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7119<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7120 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7121</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7122
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7123<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7124
7125<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7126<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
7127 floating point or vector of floating point type. Not all targets support all
7128 types however.</p>
7129
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7130<pre>
7131 declare float @llvm.pow.f32(float %Val, float %Power)
7132 declare double @llvm.pow.f64(double %Val, double %Power)
7133 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7134 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7135 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7136</pre>
7137
7138<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7139<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
7140 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7141
7142<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7143<p>The second argument is a floating point power, and the first is a value to
7144 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7145
7146<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7147<p>This function returns the first value raised to the second power, returning
7148 the same values as the libm <tt>pow</tt> functions would, and handles error
7149 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7150
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7151</div>
7152
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7153<!-- _______________________________________________________________________ -->
7154<h4>
7155 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
7156</h4>
7157
7158<div>
7159
7160<h5>Syntax:</h5>
7161<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
7162 floating point or vector of floating point type. Not all targets support all
7163 types however.</p>
7164
7165<pre>
7166 declare float @llvm.exp.f32(float %Val)
7167 declare double @llvm.exp.f64(double %Val)
7168 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7169 declare fp128 @llvm.exp.f128(fp128 %Val)
7170 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7171</pre>
7172
7173<h5>Overview:</h5>
7174<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
7175
7176<h5>Arguments:</h5>
7177<p>The argument and return value are floating point numbers of the same
7178 type.</p>
7179
7180<h5>Semantics:</h5>
7181<p>This function returns the same values as the libm <tt>exp</tt> functions
7182 would, and handles error conditions in the same way.</p>
7183
7184</div>
7185
7186<!-- _______________________________________________________________________ -->
7187<h4>
7188 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
7189</h4>
7190
7191<div>
7192
7193<h5>Syntax:</h5>
7194<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
7195 floating point or vector of floating point type. Not all targets support all
7196 types however.</p>
7197
7198<pre>
7199 declare float @llvm.log.f32(float %Val)
7200 declare double @llvm.log.f64(double %Val)
7201 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7202 declare fp128 @llvm.log.f128(fp128 %Val)
7203 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7204</pre>
7205
7206<h5>Overview:</h5>
7207<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
7208
7209<h5>Arguments:</h5>
7210<p>The argument and return value are floating point numbers of the same
7211 type.</p>
7212
7213<h5>Semantics:</h5>
7214<p>This function returns the same values as the libm <tt>log</tt> functions
7215 would, and handles error conditions in the same way.</p>
7216
Nick Lewycky1c929be2011-10-31 01:32:21 +0000[diff] [blame]7217</div>
7218
7219<!-- _______________________________________________________________________ -->
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7220<h4>
7221 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
7222</h4>
7223
7224<div>
7225
7226<h5>Syntax:</h5>
7227<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
7228 floating point or vector of floating point type. Not all targets support all
7229 types however.</p>
7230
7231<pre>
7232 declare float @llvm.fma.f32(float %a, float %b, float %c)
7233 declare double @llvm.fma.f64(double %a, double %b, double %c)
7234 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7235 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7236 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7237</pre>
7238
7239<h5>Overview:</h5>
Cameron Zwarichabc43e62011-07-08 22:13:55 +0000[diff] [blame]7240<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7241 operation.</p>
7242
7243<h5>Arguments:</h5>
7244<p>The argument and return value are floating point numbers of the same
7245 type.</p>
7246
7247<h5>Semantics:</h5>
7248<p>This function returns the same values as the libm <tt>fma</tt> functions
7249 would.</p>
7250
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7251</div>
7252
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]7253</div>
7254
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7255<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7256<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7257 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7258</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7259
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7260<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7261
7262<p>LLVM provides intrinsics for a few important bit manipulation operations.
7263 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7264
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7265<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7266<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7267 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7268</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7269
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7270<div>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7271
7272<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7273<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7274 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
7275
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7276<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7277 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
7278 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
7279 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7280</pre>
7281
7282<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7283<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
7284 values with an even number of bytes (positive multiple of 16 bits). These
7285 are useful for performing operations on data that is not in the target's
7286 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7287
7288<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7289<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
7290 and low byte of the input i16 swapped. Similarly,
7291 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
7292 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7293 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7294 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7295 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7296 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7297
7298</div>
7299
7300<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7301<h4>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]7302 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7303</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7304
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7305<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7306
7307<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7308<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]7309 width, or on any vector with integer elements. Not all targets support all
7310 bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7311
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7312<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7313 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7314 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]7315 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7316 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
7317 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7318 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]7319</pre>
7320
7321<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7322<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7323 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7324
7325<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7326<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]7327 integer type, or a vector with integer elements.
7328 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7329
7330<h5>Semantics:</h5>
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7331<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7332 element of a vector.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7333
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7334</div>
7335
7336<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7337<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7338 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7339</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7340
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7341<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7342
7343<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7344<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]7345 integer bit width, or any vector whose elements are integers. Not all
7346 targets support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7347
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7348<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7349 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7350 declare i16 @llvm.ctlz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7351 declare i32 @llvm.ctlz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7352 declare i64 @llvm.ctlz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7353 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7354 declase &lt;2 x i32&gt; @llvm.ctlz.v2i32(&lt;2 x i32&gt; &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7355</pre>
7356
7357<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7358<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7359 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7360
7361<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7362<p>The first argument is the value to be counted. This argument may be of any
7363 integer type, or a vectory with integer element type. The return type
7364 must match the first argument type.</p>
7365
7366<p>The second argument must be a constant and is a flag to indicate whether the
7367 intrinsic should ensure that a zero as the first argument produces a defined
7368 result. Historically some architectures did not provide a defined result for
7369 zero values as efficiently, and many algorithms are now predicated on
7370 avoiding zero-value inputs.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7371
7372<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7373<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7374 zeros in a variable, or within each element of the vector.
7375 If <tt>src == 0</tt> then the result is the size in bits of the type of
7376 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7377 For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7378
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7379</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7380
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7381<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7382<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7383 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7384</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7385
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7386<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7387
7388<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7389<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]7390 integer bit width, or any vector of integer elements. Not all targets
7391 support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7392
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7393<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7394 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7395 declare i16 @llvm.cttz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7396 declare i32 @llvm.cttz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7397 declare i64 @llvm.cttz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7398 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7399 declase &lt;2 x i32&gt; @llvm.cttz.v2i32(&lt;2 x i32&gt; &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7400</pre>
7401
7402<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7403<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7404 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7405
7406<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7407<p>The first argument is the value to be counted. This argument may be of any
7408 integer type, or a vectory with integer element type. The return type
7409 must match the first argument type.</p>
7410
7411<p>The second argument must be a constant and is a flag to indicate whether the
7412 intrinsic should ensure that a zero as the first argument produces a defined
7413 result. Historically some architectures did not provide a defined result for
7414 zero values as efficiently, and many algorithms are now predicated on
7415 avoiding zero-value inputs.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7416
7417<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7418<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7419 zeros in a variable, or within each element of a vector.
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7420 If <tt>src == 0</tt> then the result is the size in bits of the type of
7421 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7422 For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7423
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7424</div>
7425
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7426</div>
7427
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7428<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7429<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7430 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7431</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7432
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7433<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7434
7435<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7436
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7437<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7438<h4>
7439 <a name="int_sadd_overflow">
7440 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7441 </a>
7442</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7443
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7444<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7445
7446<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7447<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]7448 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7449
7450<pre>
7451 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7452 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7453 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7454</pre>
7455
7456<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7457<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7458 a signed addition of the two arguments, and indicate whether an overflow
7459 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7460
7461<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7462<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]7463 be of integer types of any bit width, but they must have the same bit
7464 width. The second element of the result structure must be of
7465 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7466 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7467
7468<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7469<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7470 a signed addition of the two variables. They return a structure &mdash; the
7471 first element of which is the signed summation, and the second element of
7472 which is a bit specifying if the signed summation resulted in an
7473 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7474
7475<h5>Examples:</h5>
7476<pre>
7477 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7478 %sum = extractvalue {i32, i1} %res, 0
7479 %obit = extractvalue {i32, i1} %res, 1
7480 br i1 %obit, label %overflow, label %normal
7481</pre>
7482
7483</div>
7484
7485<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7486<h4>
7487 <a name="int_uadd_overflow">
7488 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7489 </a>
7490</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7491
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7492<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7493
7494<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7495<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]7496 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7497
7498<pre>
7499 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7500 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7501 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7502</pre>
7503
7504<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7505<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7506 an unsigned addition of the two arguments, and indicate whether a carry
7507 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7508
7509<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7510<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]7511 be of integer types of any bit width, but they must have the same bit
7512 width. The second element of the result structure must be of
7513 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7514 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7515
7516<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7517<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7518 an unsigned addition of the two arguments. They return a structure &mdash;
7519 the first element of which is the sum, and the second element of which is a
7520 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7521
7522<h5>Examples:</h5>
7523<pre>
7524 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7525 %sum = extractvalue {i32, i1} %res, 0
7526 %obit = extractvalue {i32, i1} %res, 1
7527 br i1 %obit, label %carry, label %normal
7528</pre>
7529
7530</div>
7531
7532<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7533<h4>
7534 <a name="int_ssub_overflow">
7535 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7536 </a>
7537</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7538
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7539<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7540
7541<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7542<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]7543 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7544
7545<pre>
7546 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7547 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7548 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7549</pre>
7550
7551<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7552<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7553 a signed subtraction of the two arguments, and indicate whether an overflow
7554 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7555
7556<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7557<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]7558 be of integer types of any bit width, but they must have the same bit
7559 width. The second element of the result structure must be of
7560 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7561 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7562
7563<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7564<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7565 a signed subtraction of the two arguments. They return a structure &mdash;
7566 the first element of which is the subtraction, and the second element of
7567 which is a bit specifying if the signed subtraction resulted in an
7568 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7569
7570<h5>Examples:</h5>
7571<pre>
7572 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7573 %sum = extractvalue {i32, i1} %res, 0
7574 %obit = extractvalue {i32, i1} %res, 1
7575 br i1 %obit, label %overflow, label %normal
7576</pre>
7577
7578</div>
7579
7580<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7581<h4>
7582 <a name="int_usub_overflow">
7583 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7584 </a>
7585</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7586
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7587<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7588
7589<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7590<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]7591 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7592
7593<pre>
7594 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7595 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7596 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7597</pre>
7598
7599<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7600<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7601 an unsigned subtraction of the two arguments, and indicate whether an
7602 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7603
7604<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7605<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]7606 be of integer types of any bit width, but they must have the same bit
7607 width. The second element of the result structure must be of
7608 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7609 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7610
7611<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7612<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7613 an unsigned subtraction of the two arguments. They return a structure &mdash;
7614 the first element of which is the subtraction, and the second element of
7615 which is a bit specifying if the unsigned subtraction resulted in an
7616 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7617
7618<h5>Examples:</h5>
7619<pre>
7620 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7621 %sum = extractvalue {i32, i1} %res, 0
7622 %obit = extractvalue {i32, i1} %res, 1
7623 br i1 %obit, label %overflow, label %normal
7624</pre>
7625
7626</div>
7627
7628<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7629<h4>
7630 <a name="int_smul_overflow">
7631 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7632 </a>
7633</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7634
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7635<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7636
7637<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7638<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]7639 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7640
7641<pre>
7642 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7643 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7644 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7645</pre>
7646
7647<h5>Overview:</h5>
7648
7649<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7650 a signed multiplication of the two arguments, and indicate whether an
7651 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7652
7653<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7654<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]7655 be of integer types of any bit width, but they must have the same bit
7656 width. The second element of the result structure must be of
7657 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7658 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7659
7660<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7661<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7662 a signed multiplication of the two arguments. They return a structure &mdash;
7663 the first element of which is the multiplication, and the second element of
7664 which is a bit specifying if the signed multiplication resulted in an
7665 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7666
7667<h5>Examples:</h5>
7668<pre>
7669 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7670 %sum = extractvalue {i32, i1} %res, 0
7671 %obit = extractvalue {i32, i1} %res, 1
7672 br i1 %obit, label %overflow, label %normal
7673</pre>
7674
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]7675</div>
7676
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7677<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7678<h4>
7679 <a name="int_umul_overflow">
7680 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
7681 </a>
7682</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7683
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7684<div>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7685
7686<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7687<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]7688 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7689
7690<pre>
7691 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
7692 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7693 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
7694</pre>
7695
7696<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7697<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7698 a unsigned multiplication of the two arguments, and indicate whether an
7699 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7700
7701<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7702<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]7703 be of integer types of any bit width, but they must have the same bit
7704 width. The second element of the result structure must be of
7705 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7706 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7707
7708<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7709<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7710 an unsigned multiplication of the two arguments. They return a structure
7711 &mdash; the first element of which is the multiplication, and the second
7712 element of which is a bit specifying if the unsigned multiplication resulted
7713 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]7714
7715<h5>Examples:</h5>
7716<pre>
7717 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
7718 %sum = extractvalue {i32, i1} %res, 0
7719 %obit = extractvalue {i32, i1} %res, 1
7720 br i1 %obit, label %overflow, label %normal
7721</pre>
7722
7723</div>
7724
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7725</div>
7726
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7727<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7728<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7729 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7730</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7731
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7732<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7733
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7734<p>Half precision floating point is a storage-only format. This means that it is
7735 a dense encoding (in memory) but does not support computation in the
7736 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7737
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7738<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7739 value as an i16, then convert it to float with <a
7740 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
7741 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7742 double etc). To store the value back to memory, it is first converted to
7743 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7744 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
7745 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7746
7747<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7748<h4>
7749 <a name="int_convert_to_fp16">
7750 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
7751 </a>
7752</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7753
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7754<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7755
7756<h5>Syntax:</h5>
7757<pre>
7758 declare i16 @llvm.convert.to.fp16(f32 %a)
7759</pre>
7760
7761<h5>Overview:</h5>
7762<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7763 a conversion from single precision floating point format to half precision
7764 floating point format.</p>
7765
7766<h5>Arguments:</h5>
7767<p>The intrinsic function contains single argument - the value to be
7768 converted.</p>
7769
7770<h5>Semantics:</h5>
7771<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
7772 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7773 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7774 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7775
7776<h5>Examples:</h5>
7777<pre>
7778 %res = call i16 @llvm.convert.to.fp16(f32 %a)
7779 store i16 %res, i16* @x, align 2
7780</pre>
7781
7782</div>
7783
7784<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7785<h4>
7786 <a name="int_convert_from_fp16">
7787 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
7788 </a>
7789</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7790
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7791<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7792
7793<h5>Syntax:</h5>
7794<pre>
7795 declare f32 @llvm.convert.from.fp16(i16 %a)
7796</pre>
7797
7798<h5>Overview:</h5>
7799<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
7800 a conversion from half precision floating point format to single precision
7801 floating point format.</p>
7802
7803<h5>Arguments:</h5>
7804<p>The intrinsic function contains single argument - the value to be
7805 converted.</p>
7806
7807<h5>Semantics:</h5>
7808<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]7809 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]7810 precision floating point format. The input half-float value is represented by
7811 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7812
7813<h5>Examples:</h5>
7814<pre>
7815 %a = load i16* @x, align 2
7816 %res = call f32 @llvm.convert.from.fp16(i16 %a)
7817</pre>
7818
7819</div>
7820
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7821</div>
7822
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]7823<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7824<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7825 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7826</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7827
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7828<div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7829
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7830<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
7831 prefix), are described in
7832 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
7833 Level Debugging</a> document.</p>
7834
7835</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]7836
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7837<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7838<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7839 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7840</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7841
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7842<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7843
7844<p>The LLVM exception handling intrinsics (which all start with
7845 <tt>llvm.eh.</tt> prefix), are described in
7846 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
7847 Handling</a> document.</p>
7848
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]7849</div>
7850
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]7851<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7852<h3>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7853 <a name="int_trampoline">Trampoline Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7854</h3>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7855
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7856<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7857
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7858<p>These intrinsics make it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]7859 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
7860 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7861 function pointer lacking the nest parameter - the caller does not need to
7862 provide a value for it. Instead, the value to use is stored in advance in a
7863 "trampoline", a block of memory usually allocated on the stack, which also
7864 contains code to splice the nest value into the argument list. This is used
7865 to implement the GCC nested function address extension.</p>
7866
7867<p>For example, if the function is
7868 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
7869 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
7870 follows:</p>
7871
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]7872<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]7873 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
7874 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7875 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
7876 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]7877 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7878</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7879
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7880<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
7881 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7882
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7883<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7884<h4>
7885 <a name="int_it">
7886 '<tt>llvm.init.trampoline</tt>' Intrinsic
7887 </a>
7888</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7889
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7890<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7891
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7892<h5>Syntax:</h5>
7893<pre>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7894 declare void @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7895</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7896
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7897<h5>Overview:</h5>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7898<p>This fills the memory pointed to by <tt>tramp</tt> with executable code,
7899 turning it into a trampoline.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7900
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7901<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7902<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
7903 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
7904 sufficiently aligned block of memory; this memory is written to by the
7905 intrinsic. Note that the size and the alignment are target-specific - LLVM
7906 currently provides no portable way of determining them, so a front-end that
7907 generates this intrinsic needs to have some target-specific knowledge.
7908 The <tt>func</tt> argument must hold a function bitcast to
7909 an <tt>i8*</tt>.</p>
7910
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7911<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7912<p>The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]7913 dependent code, turning it into a function. Then <tt>tramp</tt> needs to be
7914 passed to <a href="#int_at">llvm.adjust.trampoline</a> to get a pointer
7915 which can be <a href="#int_trampoline">bitcast (to a new function) and
7916 called</a>. The new function's signature is the same as that of
7917 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
7918 removed. At most one such <tt>nest</tt> argument is allowed, and it must be of
7919 pointer type. Calling the new function is equivalent to calling <tt>func</tt>
7920 with the same argument list, but with <tt>nval</tt> used for the missing
7921 <tt>nest</tt> argument. If, after calling <tt>llvm.init.trampoline</tt>, the
7922 memory pointed to by <tt>tramp</tt> is modified, then the effect of any later call
7923 to the returned function pointer is undefined.</p>
7924</div>
7925
7926<!-- _______________________________________________________________________ -->
7927<h4>
7928 <a name="int_at">
7929 '<tt>llvm.adjust.trampoline</tt>' Intrinsic
7930 </a>
7931</h4>
7932
7933<div>
7934
7935<h5>Syntax:</h5>
7936<pre>
7937 declare i8* @llvm.adjust.trampoline(i8* &lt;tramp&gt;)
7938</pre>
7939
7940<h5>Overview:</h5>
7941<p>This performs any required machine-specific adjustment to the address of a
7942 trampoline (passed as <tt>tramp</tt>).</p>
7943
7944<h5>Arguments:</h5>
7945<p><tt>tramp</tt> must point to a block of memory which already has trampoline code
7946 filled in by a previous call to <a href="#int_it"><tt>llvm.init.trampoline</tt>
7947 </a>.</p>
7948
7949<h5>Semantics:</h5>
7950<p>On some architectures the address of the code to be executed needs to be
7951 different to the address where the trampoline is actually stored. This
7952 intrinsic returns the executable address corresponding to <tt>tramp</tt>
7953 after performing the required machine specific adjustments.
7954 The pointer returned can then be <a href="#int_trampoline"> bitcast and
7955 executed</a>.
7956</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7957
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7958</div>
7959
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7960</div>
7961
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]7962<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7963<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7964 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7965</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7966
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7967<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7968
7969<p>This class of intrinsics exists to information about the lifetime of memory
7970 objects and ranges where variables are immutable.</p>
7971
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7972<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7973<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7974 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7975</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7976
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7977<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7978
7979<h5>Syntax:</h5>
7980<pre>
7981 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
7982</pre>
7983
7984<h5>Overview:</h5>
7985<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
7986 object's lifetime.</p>
7987
7988<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]7989<p>The first argument is a constant integer representing the size of the
7990 object, or -1 if it is variable sized. The second argument is a pointer to
7991 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7992
7993<h5>Semantics:</h5>
7994<p>This intrinsic indicates that before this point in the code, the value of the
7995 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]7996 never be used and has an undefined value. A load from the pointer that
7997 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]7998 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
7999
8000</div>
8001
8002<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8003<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8004 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8005</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8006
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8007<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8008
8009<h5>Syntax:</h5>
8010<pre>
8011 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8012</pre>
8013
8014<h5>Overview:</h5>
8015<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
8016 object's lifetime.</p>
8017
8018<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8019<p>The first argument is a constant integer representing the size of the
8020 object, or -1 if it is variable sized. The second argument is a pointer to
8021 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8022
8023<h5>Semantics:</h5>
8024<p>This intrinsic indicates that after this point in the code, the value of the
8025 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8026 never be used and has an undefined value. Any stores into the memory object
8027 following this intrinsic may be removed as dead.
8028
8029</div>
8030
8031<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8032<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8033 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8034</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8035
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8036<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8037
8038<h5>Syntax:</h5>
8039<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +0000[diff] [blame]8040 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8041</pre>
8042
8043<h5>Overview:</h5>
8044<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8045 a memory object will not change.</p>
8046
8047<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8048<p>The first argument is a constant integer representing the size of the
8049 object, or -1 if it is variable sized. The second argument is a pointer to
8050 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8051
8052<h5>Semantics:</h5>
8053<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8054 the return value, the referenced memory location is constant and
8055 unchanging.</p>
8056
8057</div>
8058
8059<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8060<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8061 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8062</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8063
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8064<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8065
8066<h5>Syntax:</h5>
8067<pre>
8068 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8069</pre>
8070
8071<h5>Overview:</h5>
8072<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8073 a memory object are mutable.</p>
8074
8075<h5>Arguments:</h5>
8076<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8077 The second argument is a constant integer representing the size of the
8078 object, or -1 if it is variable sized and the third argument is a pointer
8079 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8080
8081<h5>Semantics:</h5>
8082<p>This intrinsic indicates that the memory is mutable again.</p>
8083
8084</div>
8085
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8086</div>
8087
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]8088<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8089<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8090 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8091</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8092
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8093<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8094
8095<p>This class of intrinsics is designed to be generic and has no specific
8096 purpose.</p>
8097
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8098<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8099<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8100 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8101</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8102
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8103<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8104
8105<h5>Syntax:</h5>
8106<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8107 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]8108</pre>
8109
8110<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8111<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8112
8113<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8114<p>The first argument is a pointer to a value, the second is a pointer to a
8115 global string, the third is a pointer to a global string which is the source
8116 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8117
8118<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8119<p>This intrinsic allows annotation of local variables with arbitrary strings.
8120 This can be useful for special purpose optimizations that want to look for
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8121 these annotations. These have no other defined use; they are ignored by code
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8122 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8123
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8124</div>
8125
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8126<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8127<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]8128 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8129</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8130
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8131<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8132
8133<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8134<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8135 any integer bit width.</p>
8136
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8137<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8138 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8139 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8140 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8141 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8142 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]8143</pre>
8144
8145<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8146<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8147
8148<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8149<p>The first argument is an integer value (result of some expression), the
8150 second is a pointer to a global string, the third is a pointer to a global
8151 string which is the source file name, and the last argument is the line
8152 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8153
8154<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8155<p>This intrinsic allows annotations to be put on arbitrary expressions with
8156 arbitrary strings. This can be useful for special purpose optimizations that
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8157 want to look for these annotations. These have no other defined use; they
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8158 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8159
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8160</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8161
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8162<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8163<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8164 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8165</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8166
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8167<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8168
8169<h5>Syntax:</h5>
8170<pre>
8171 declare void @llvm.trap()
8172</pre>
8173
8174<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8175<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8176
8177<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8178<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8179
8180<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8181<p>This intrinsics is lowered to the target dependent trap instruction. If the
8182 target does not have a trap instruction, this intrinsic will be lowered to
8183 the call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8184
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8185</div>
8186
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8187<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8188<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +0000[diff] [blame]8189 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8190</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8191
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8192<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8193
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8194<h5>Syntax:</h5>
8195<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8196 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8197</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8198
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8199<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8200<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8201 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8202 ensure that it is placed on the stack before local variables.</p>
8203
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8204<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8205<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8206 arguments. The first argument is the value loaded from the stack
8207 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8208 that has enough space to hold the value of the guard.</p>
8209
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8210<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8211<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8212 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8213 stack. This is to ensure that if a local variable on the stack is
8214 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8215 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8216 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8217 function.</p>
8218
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8219</div>
8220
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8221<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8222<h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8223 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8224</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8225
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8226<div>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8227
8228<h5>Syntax:</h5>
8229<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8230 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;type&gt;)
8231 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;type&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8232</pre>
8233
8234<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8235<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8236 the optimizers to determine at compile time whether a) an operation (like
8237 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8238 runtime check for overflow isn't necessary. An object in this context means
8239 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8240
8241<h5>Arguments:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8242<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8243 argument is a pointer to or into the <tt>object</tt>. The second argument
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8244 is a boolean 0 or 1. This argument determines whether you want the
8245 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]8246 1, variables are not allowed.</p>
8247
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8248<h5>Semantics:</h5>
8249<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to either a constant
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8250 representing the size of the object concerned, or <tt>i32/i64 -1 or 0</tt>,
8251 depending on the <tt>type</tt> argument, if the size cannot be determined at
8252 compile time.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8253
8254</div>
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8255<!-- _______________________________________________________________________ -->
8256<h4>
8257 <a name="int_expect">'<tt>llvm.expect</tt>' Intrinsic</a>
8258</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8259
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8260<div>
8261
8262<h5>Syntax:</h5>
8263<pre>
8264 declare i32 @llvm.expect.i32(i32 &lt;val&gt;, i32 &lt;expected_val&gt;)
8265 declare i64 @llvm.expect.i64(i64 &lt;val&gt;, i64 &lt;expected_val&gt;)
8266</pre>
8267
8268<h5>Overview:</h5>
8269<p>The <tt>llvm.expect</tt> intrinsic provides information about expected (the
8270 most probable) value of <tt>val</tt>, which can be used by optimizers.</p>
8271
8272<h5>Arguments:</h5>
8273<p>The <tt>llvm.expect</tt> intrinsic takes two arguments. The first
8274 argument is a value. The second argument is an expected value, this needs to
8275 be a constant value, variables are not allowed.</p>
8276
8277<h5>Semantics:</h5>
8278<p>This intrinsic is lowered to the <tt>val</tt>.</p>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8279</div>
8280
8281</div>
8282
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8283</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8284<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8285<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8286<address>
8287 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8288 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8289 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8290 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]8291
8292 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]8293 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8294 Last modified: $Date$
8295</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]8296
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]8297</body>
8298</html>