Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 7bebddf55ece46995f310d79195afb4e5b239886 / . / docs / LangRef.html
blob: 0699857bb1e45aef203615f85b606170a933bb0c [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.">
Daniel Dunbaradea4972012-04-19 20:20:34 +0000[diff] [blame]10 <link rel="stylesheet" href="_static/llvm.css" type="text/css">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]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 Wendling987e7eb2009-07-20 02:41:50 +0000[diff] [blame]28 <li><a href="#linkage_internal">'<tt>internal</tt>' Linkage</a></li>
29 <li><a href="#linkage_available_externally">'<tt>available_externally</tt>' Linkage</a></li>
30 <li><a href="#linkage_linkonce">'<tt>linkonce</tt>' Linkage</a></li>
31 <li><a href="#linkage_common">'<tt>common</tt>' Linkage</a></li>
32 <li><a href="#linkage_weak">'<tt>weak</tt>' Linkage</a></li>
33 <li><a href="#linkage_appending">'<tt>appending</tt>' Linkage</a></li>
34 <li><a href="#linkage_externweak">'<tt>extern_weak</tt>' Linkage</a></li>
Chris Lattner5a2d8752009-10-10 18:26:06 +0000[diff] [blame]35 <li><a href="#linkage_linkonce_odr">'<tt>linkonce_odr</tt>' Linkage</a></li>
Bill Wendling32811be2012-08-17 18:33:14 +0000[diff] [blame]36 <li><a href="#linkage_linkonce_odr_auto_hide">'<tt>linkonce_odr_auto_hide</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>
Duncan Sands5e5c5f82012-04-14 12:36:06 +0000[diff] [blame]106 <li><a href="#fpmath">'<tt>fpmath</tt>' Metadata</a></li>
Rafael Espindola39dd3282012-03-24 00:14:51 +0000[diff] [blame]107 <li><a href="#range">'<tt>range</tt>' Metadata</a></li>
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]108 </ol>
109 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]110 </ol>
111 </li>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]112 <li><a href="#module_flags">Module Flags Metadata</a>
113 <ol>
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]114 <li><a href="#objc_gc_flags">Objective-C Garbage Collection Module Flags Metadata</a></li>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]115 </ol>
116 </li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]117 <li><a href="#intrinsic_globals">Intrinsic Global Variables</a>
118 <ol>
119 <li><a href="#intg_used">The '<tt>llvm.used</tt>' Global Variable</a></li>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]120 <li><a href="#intg_compiler_used">The '<tt>llvm.compiler.used</tt>'
121 Global Variable</a></li>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]122 <li><a href="#intg_global_ctors">The '<tt>llvm.global_ctors</tt>'
123 Global Variable</a></li>
124 <li><a href="#intg_global_dtors">The '<tt>llvm.global_dtors</tt>'
125 Global Variable</a></li>
126 </ol>
127 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]128 <li><a href="#instref">Instruction Reference</a>
129 <ol>
130 <li><a href="#terminators">Terminator Instructions</a>
131 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]132 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
133 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]134 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]135 <li><a href="#i_indirectbr">'<tt>indirectbr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]136 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]137 <li><a href="#i_resume">'<tt>resume</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]138 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]139 </ol>
140 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]141 <li><a href="#binaryops">Binary Operations</a>
142 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]143 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]144 <li><a href="#i_fadd">'<tt>fadd</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]145 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]146 <li><a href="#i_fsub">'<tt>fsub</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]147 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]148 <li><a href="#i_fmul">'<tt>fmul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]149 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
150 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
151 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]152 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
153 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
154 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]155 </ol>
156 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]157 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
158 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]159 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
160 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
161 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]162 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]163 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]164 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]165 </ol>
166 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]167 <li><a href="#vectorops">Vector Operations</a>
168 <ol>
169 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
170 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
171 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]172 </ol>
173 </li>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]174 <li><a href="#aggregateops">Aggregate Operations</a>
175 <ol>
176 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
177 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
178 </ol>
179 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]180 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]181 <ol>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]182 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
183 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
184 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
185 <li><a href="#i_fence">'<tt>fence</tt>' Instruction</a></li>
186 <li><a href="#i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a></li>
187 <li><a href="#i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]188 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]189 </ol>
190 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]191 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]192 <ol>
193 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
194 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
195 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
196 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
197 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]198 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
199 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
200 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
201 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]202 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
203 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]204 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]205 </ol>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]206 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]207 <li><a href="#otherops">Other Operations</a>
208 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]209 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
210 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]211 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]212 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]213 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]214 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]215 <li><a href="#i_landingpad">'<tt>landingpad</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]216 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]217 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]218 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]219 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]220 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]221 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]222 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
223 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]224 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
225 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
226 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]227 </ol>
228 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]229 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
230 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]231 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
232 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
233 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]234 </ol>
235 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]236 <li><a href="#int_codegen">Code Generator Intrinsics</a>
237 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]238 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
239 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
240 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
241 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
242 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
243 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Dan Gohman31f1af12010-05-26 21:56:15 +0000[diff] [blame]244 <li><a href="#int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]245 </ol>
246 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]247 <li><a href="#int_libc">Standard C Library Intrinsics</a>
248 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]249 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
250 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
251 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
252 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
253 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]254 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
255 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
256 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Dan Gohman08b280b2011-05-27 00:36:31 +0000[diff] [blame]257 <li><a href="#int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a></li>
258 <li><a href="#int_log">'<tt>llvm.log.*</tt>' Intrinsic</a></li>
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]259 <li><a href="#int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a></li>
Peter Collingbourne168a4c32012-07-03 12:25:40 +0000[diff] [blame]260 <li><a href="#int_fabs">'<tt>llvm.fabs.*</tt>' Intrinsic</a></li>
Dan Gohman27db99f2012-07-26 17:43:27 +0000[diff] [blame]261 <li><a href="#int_floor">'<tt>llvm.floor.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]262 </ol>
263 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]264 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]265 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]266 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]267 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
268 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
269 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]270 </ol>
271 </li>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]272 <li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
273 <ol>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]274 <li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
275 <li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
276 <li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
277 <li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
278 <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]279 <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]280 </ol>
281 </li>
Lang Hames5afba6f2012-06-05 19:07:46 +0000[diff] [blame]282 <li><a href="#spec_arithmetic">Specialised Arithmetic Intrinsics</a>
283 <ol>
284 <li><a href="#fmuladd">'<tt>llvm.fmuladd</tt> Intrinsic</a></li>
285 </ol>
286 </li>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]287 <li><a href="#int_fp16">Half Precision Floating Point Intrinsics</a>
288 <ol>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]289 <li><a href="#int_convert_to_fp16">'<tt>llvm.convert.to.fp16</tt>' Intrinsic</a></li>
290 <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]291 </ol>
292 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]293 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]294 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]295 <li><a href="#int_trampoline">Trampoline Intrinsics</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]296 <ol>
297 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]298 <li><a href="#int_at">'<tt>llvm.adjust.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]299 </ol>
300 </li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]301 <li><a href="#int_memorymarkers">Memory Use Markers</a>
302 <ol>
Jakub Staszak8e1b12a2011-12-04 20:44:25 +0000[diff] [blame]303 <li><a href="#int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a></li>
304 <li><a href="#int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a></li>
305 <li><a href="#int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a></li>
306 <li><a href="#int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a></li>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]307 </ol>
308 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]309 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]310 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]311 <li><a href="#int_var_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]312 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]313 <li><a href="#int_annotation">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]314 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]315 <li><a href="#int_trap">
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]316 '<tt>llvm.trap</tt>' Intrinsic</a></li>
Dan Gohmana6063c62012-05-14 18:58:10 +0000[diff] [blame]317 <li><a href="#int_debugtrap">
318 '<tt>llvm.debugtrap</tt>' Intrinsic</a></li>
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]319 <li><a href="#int_stackprotector">
320 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Nuno Lopesb0c76d72012-07-05 17:37:07 +0000[diff] [blame]321 <li><a href="#int_objectsize">
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]322 '<tt>llvm.objectsize</tt>' Intrinsic</a></li>
Nuno Lopesb0c76d72012-07-05 17:37:07 +0000[diff] [blame]323 <li><a href="#int_expect">
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]324 '<tt>llvm.expect</tt>' Intrinsic</a></li>
Nuno Lopesb0c76d72012-07-05 17:37:07 +0000[diff] [blame]325 <li><a href="#int_donothing">
326 '<tt>llvm.donothing</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]327 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]328 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]329 </ol>
330 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]331</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]332
333<div class="doc_author">
334 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
335 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]336</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]337
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]338<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]339<h2><a name="abstract">Abstract</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]340<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]341
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]342<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]343
344<p>This document is a reference manual for the LLVM assembly language. LLVM is
345 a Static Single Assignment (SSA) based representation that provides type
346 safety, low-level operations, flexibility, and the capability of representing
347 'all' high-level languages cleanly. It is the common code representation
348 used throughout all phases of the LLVM compilation strategy.</p>
349
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]350</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]351
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]352<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]353<h2><a name="introduction">Introduction</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]354<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]355
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]356<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]357
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]358<p>The LLVM code representation is designed to be used in three different forms:
359 as an in-memory compiler IR, as an on-disk bitcode representation (suitable
360 for fast loading by a Just-In-Time compiler), and as a human readable
361 assembly language representation. This allows LLVM to provide a powerful
362 intermediate representation for efficient compiler transformations and
363 analysis, while providing a natural means to debug and visualize the
364 transformations. The three different forms of LLVM are all equivalent. This
365 document describes the human readable representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]366
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]367<p>The LLVM representation aims to be light-weight and low-level while being
368 expressive, typed, and extensible at the same time. It aims to be a
369 "universal IR" of sorts, by being at a low enough level that high-level ideas
370 may be cleanly mapped to it (similar to how microprocessors are "universal
371 IR's", allowing many source languages to be mapped to them). By providing
372 type information, LLVM can be used as the target of optimizations: for
373 example, through pointer analysis, it can be proven that a C automatic
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]374 variable is never accessed outside of the current function, allowing it to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]375 be promoted to a simple SSA value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]376
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]377<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]378<h4>
379 <a name="wellformed">Well-Formedness</a>
380</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]381
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]382<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]383
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]384<p>It is important to note that this document describes 'well formed' LLVM
385 assembly language. There is a difference between what the parser accepts and
386 what is considered 'well formed'. For example, the following instruction is
387 syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]388
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]389<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]390%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]391</pre>
392
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]393<p>because the definition of <tt>%x</tt> does not dominate all of its uses. The
394 LLVM infrastructure provides a verification pass that may be used to verify
395 that an LLVM module is well formed. This pass is automatically run by the
396 parser after parsing input assembly and by the optimizer before it outputs
397 bitcode. The violations pointed out by the verifier pass indicate bugs in
398 transformation passes or input to the parser.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]399
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]400</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]401
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]402</div>
403
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame]404<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]405
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]406<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]407<h2><a name="identifiers">Identifiers</a></h2>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]408<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]409
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]410<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]411
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]412<p>LLVM identifiers come in two basic types: global and local. Global
413 identifiers (functions, global variables) begin with the <tt>'@'</tt>
414 character. Local identifiers (register names, types) begin with
415 the <tt>'%'</tt> character. Additionally, there are three different formats
416 for identifiers, for different purposes:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]417
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]418<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]419 <li>Named values are represented as a string of characters with their prefix.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]420 For example, <tt>%foo</tt>, <tt>@DivisionByZero</tt>,
421 <tt>%a.really.long.identifier</tt>. The actual regular expression used is
422 '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'. Identifiers which require
423 other characters in their names can be surrounded with quotes. Special
424 characters may be escaped using <tt>"\xx"</tt> where <tt>xx</tt> is the
425 ASCII code for the character in hexadecimal. In this way, any character
426 can be used in a name value, even quotes themselves.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]427
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]428 <li>Unnamed values are represented as an unsigned numeric value with their
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]429 prefix. For example, <tt>%12</tt>, <tt>@2</tt>, <tt>%44</tt>.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]430
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]431 <li>Constants, which are described in a <a href="#constants">section about
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]432 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]433</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]434
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]435<p>LLVM requires that values start with a prefix for two reasons: Compilers
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]436 don't need to worry about name clashes with reserved words, and the set of
437 reserved words may be expanded in the future without penalty. Additionally,
438 unnamed identifiers allow a compiler to quickly come up with a temporary
439 variable without having to avoid symbol table conflicts.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]440
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]441<p>Reserved words in LLVM are very similar to reserved words in other
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]442 languages. There are keywords for different opcodes
443 ('<tt><a href="#i_add">add</a></tt>',
444 '<tt><a href="#i_bitcast">bitcast</a></tt>',
445 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names
446 ('<tt><a href="#t_void">void</a></tt>',
447 '<tt><a href="#t_primitive">i32</a></tt>', etc...), and others. These
448 reserved words cannot conflict with variable names, because none of them
449 start with a prefix character (<tt>'%'</tt> or <tt>'@'</tt>).</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]450
451<p>Here is an example of LLVM code to multiply the integer variable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]452 '<tt>%X</tt>' by 8:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]453
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]454<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]455
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]456<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]457%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]458</pre>
459
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]460<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]461
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]462<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]463%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]464</pre>
465
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]466<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]467
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]468<pre class="doc_code">
Gabor Greifec58f752009-10-28 13:05:07 +0000[diff] [blame]469%0 = <a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
470%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]471%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]472</pre>
473
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]474<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several important
475 lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]476
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]477<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]478 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]479 line.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]480
481 <li>Unnamed temporaries are created when the result of a computation is not
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]482 assigned to a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]483
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]484 <li>Unnamed temporaries are numbered sequentially</li>
485</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]486
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]487<p>It also shows a convention that we follow in this document. When
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]488 demonstrating instructions, we will follow an instruction with a comment that
489 defines the type and name of value produced. Comments are shown in italic
490 text.</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]491
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]492</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]493
494<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]495<h2><a name="highlevel">High Level Structure</a></h2>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]496<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]497<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]498<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]499<h3>
500 <a name="modulestructure">Module Structure</a>
501</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]502
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]503<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]504
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]505<p>LLVM programs are composed of <tt>Module</tt>s, each of which is a
506 translation unit of the input programs. Each module consists of functions,
507 global variables, and symbol table entries. Modules may be combined together
508 with the LLVM linker, which merges function (and global variable)
509 definitions, resolves forward declarations, and merges symbol table
510 entries. Here is an example of the "hello world" module:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]511
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]512<pre class="doc_code">
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]513<i>; Declare the string constant as a global constant.</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]514<a href="#identifiers">@.str</a> = <a href="#linkage_private">private</a>&nbsp;<a href="#globalvars">unnamed_addr</a>&nbsp;<a href="#globalvars">constant</a>&nbsp;<a href="#t_array">[13 x i8]</a> c"hello world\0A\00"&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]515
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]516<i>; External declaration of the puts function</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]517<a href="#functionstructure">declare</a> i32 @puts(i8* <a href="#nocapture">nocapture</a>) <a href="#fnattrs">nounwind</a>&nbsp;
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]518
519<i>; Definition of main function</i>
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]520define i32 @main() { <i>; i32()* </i>&nbsp;
521 <i>; Convert [13 x i8]* to i8 *...</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]522 %cast210 = <a href="#i_getelementptr">getelementptr</a> [13 x i8]* @.str, i64 0, i64 0
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]523
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]524 <i>; Call puts function to write out the string to stdout.</i>&nbsp;
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]525 <a href="#i_call">call</a> i32 @puts(i8* %cast210)
Chris Lattner63e4ccb2010-08-17 17:13:42 +0000[diff] [blame]526 <a href="#i_ret">ret</a> i32 0&nbsp;
527}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]528
529<i>; Named metadata</i>
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]530!1 = metadata !{i32 42}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]531!foo = !{!1, null}
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]532</pre>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]533
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]534<p>This example is made up of a <a href="#globalvars">global variable</a> named
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]535 "<tt>.str</tt>", an external declaration of the "<tt>puts</tt>" function,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]536 a <a href="#functionstructure">function definition</a> for
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]537 "<tt>main</tt>" and <a href="#namedmetadatastructure">named metadata</a>
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]538 "<tt>foo</tt>".</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]539
Bill Wendling4cc2be62012-03-14 08:07:43 +0000[diff] [blame]540<p>In general, a module is made up of a list of global values (where both
541 functions and global variables are global values). Global values are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]542 represented by a pointer to a memory location (in this case, a pointer to an
543 array of char, and a pointer to a function), and have one of the
544 following <a href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]545
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]546</div>
547
548<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]549<h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]550 <a name="linkage">Linkage Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]551</h3>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]552
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]553<div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]554
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]555<p>All Global Variables and Functions have one of the following types of
556 linkage:</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]557
558<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]559 <dt><tt><b><a name="linkage_private">private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]560 <dd>Global values with "<tt>private</tt>" linkage are only directly accessible
561 by objects in the current module. In particular, linking code into a
562 module with an private global value may cause the private to be renamed as
563 necessary to avoid collisions. Because the symbol is private to the
564 module, all references can be updated. This doesn't show up in any symbol
565 table in the object file.</dd>
Rafael Espindolabb46f522009-01-15 20:18:42 +0000[diff] [blame]566
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]567 <dt><tt><b><a name="linkage_linker_private">linker_private</a></b></tt></dt>
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]568 <dd>Similar to <tt>private</tt>, but the symbol is passed through the
569 assembler and evaluated by the linker. Unlike normal strong symbols, they
570 are removed by the linker from the final linked image (executable or
571 dynamic library).</dd>
572
573 <dt><tt><b><a name="linkage_linker_private_weak">linker_private_weak</a></b></tt></dt>
574 <dd>Similar to "<tt>linker_private</tt>", but the symbol is weak. Note that
575 <tt>linker_private_weak</tt> symbols are subject to coalescing by the
576 linker. The symbols are removed by the linker from the final linked image
577 (executable or dynamic library).</dd>
Bill Wendling3d10a5a2009-07-20 01:03:30 +0000[diff] [blame]578
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]579 <dt><tt><b><a name="linkage_internal">internal</a></b></tt></dt>
Bill Wendling07d31772010-06-29 22:34:52 +0000[diff] [blame]580 <dd>Similar to private, but the value shows as a local symbol
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]581 (<tt>STB_LOCAL</tt> in the case of ELF) in the object file. This
582 corresponds to the notion of the '<tt>static</tt>' keyword in C.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]583
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]584 <dt><tt><b><a name="linkage_available_externally">available_externally</a></b></tt></dt>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]585 <dd>Globals with "<tt>available_externally</tt>" linkage are never emitted
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]586 into the object file corresponding to the LLVM module. They exist to
587 allow inlining and other optimizations to take place given knowledge of
588 the definition of the global, which is known to be somewhere outside the
589 module. Globals with <tt>available_externally</tt> linkage are allowed to
590 be discarded at will, and are otherwise the same as <tt>linkonce_odr</tt>.
591 This linkage type is only allowed on definitions, not declarations.</dd>
Chris Lattner266c7bb2009-04-13 05:44:34 +0000[diff] [blame]592
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]593 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt></dt>
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]594 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
Chris Lattner873187c2010-01-09 19:15:14 +0000[diff] [blame]595 the same name when linkage occurs. This can be used to implement
596 some forms of inline functions, templates, or other code which must be
597 generated in each translation unit that uses it, but where the body may
598 be overridden with a more definitive definition later. Unreferenced
599 <tt>linkonce</tt> globals are allowed to be discarded. Note that
600 <tt>linkonce</tt> linkage does not actually allow the optimizer to
601 inline the body of this function into callers because it doesn't know if
602 this definition of the function is the definitive definition within the
603 program or whether it will be overridden by a stronger definition.
604 To enable inlining and other optimizations, use "<tt>linkonce_odr</tt>"
605 linkage.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]606
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]607 <dt><tt><b><a name="linkage_weak">weak</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]608 <dd>"<tt>weak</tt>" linkage has the same merging semantics as
609 <tt>linkonce</tt> linkage, except that unreferenced globals with
610 <tt>weak</tt> linkage may not be discarded. This is used for globals that
611 are declared "weak" in C source code.</dd>
612
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]613 <dt><tt><b><a name="linkage_common">common</a></b></tt></dt>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]614 <dd>"<tt>common</tt>" linkage is most similar to "<tt>weak</tt>" linkage, but
615 they are used for tentative definitions in C, such as "<tt>int X;</tt>" at
616 global scope.
617 Symbols with "<tt>common</tt>" linkage are merged in the same way as
618 <tt>weak symbols</tt>, and they may not be deleted if unreferenced.
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]619 <tt>common</tt> symbols may not have an explicit section,
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]620 must have a zero initializer, and may not be marked '<a
Chris Lattnercd81f5d2009-08-05 05:41:44 +0000[diff] [blame]621 href="#globalvars"><tt>constant</tt></a>'. Functions and aliases may not
622 have common linkage.</dd>
Chris Lattner26d054d2009-08-05 05:21:07 +0000[diff] [blame]623
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_appending">appending</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]626 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]627 pointer to array type. When two global variables with appending linkage
628 are linked together, the two global arrays are appended together. This is
629 the LLVM, typesafe, equivalent of having the system linker append together
630 "sections" with identical names when .o files are linked.</dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]631
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]632 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]633 <dd>The semantics of this linkage follow the ELF object file model: the symbol
634 is weak until linked, if not linked, the symbol becomes null instead of
635 being an undefined reference.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]636
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]637 <dt><tt><b><a name="linkage_linkonce_odr">linkonce_odr</a></b></tt></dt>
638 <dt><tt><b><a name="linkage_weak_odr">weak_odr</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]639 <dd>Some languages allow differing globals to be merged, such as two functions
640 with different semantics. Other languages, such as <tt>C++</tt>, ensure
Bill Wendling5e721d72010-07-01 21:55:59 +0000[diff] [blame]641 that only equivalent globals are ever merged (the "one definition rule"
642 &mdash; "ODR"). Such languages can use the <tt>linkonce_odr</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]643 and <tt>weak_odr</tt> linkage types to indicate that the global will only
644 be merged with equivalent globals. These linkage types are otherwise the
645 same as their non-<tt>odr</tt> versions.</dd>
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]646
Bill Wendling32811be2012-08-17 18:33:14 +0000[diff] [blame]647 <dt><tt><b><a name="linkage_linkonce_odr_auto_hide">linkonce_odr_auto_hide</a></b></tt></dt>
648 <dd>Similar to "<tt>linkonce_odr</tt>", but nothing in the translation unit
649 takes the address of this definition. For instance, functions that had an
650 inline definition, but the compiler decided not to inline it.
651 <tt>linkonce_odr_auto_hide</tt> may have only <tt>default</tt> visibility.
652 The symbols are removed by the linker from the final linked image
653 (executable or dynamic library).</dd>
654
Bill Wendling5c3a9f72011-11-04 20:40:41 +0000[diff] [blame]655 <dt><tt><b><a name="linkage_external">external</a></b></tt></dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]656 <dd>If none of the above identifiers are used, the global is externally
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]657 visible, meaning that it participates in linkage and can be used to
658 resolve external symbol references.</dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]659</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]660
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]661<p>The next two types of linkage are targeted for Microsoft Windows platform
662 only. They are designed to support importing (exporting) symbols from (to)
663 DLLs (Dynamic Link Libraries).</p>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]664
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]665<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]666 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]667 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]668 or variable via a global pointer to a pointer that is set up by the DLL
669 exporting the symbol. On Microsoft Windows targets, the pointer name is
670 formed by combining <code>__imp_</code> and the function or variable
671 name.</dd>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]672
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]673 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt></dt>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]674 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]675 pointer to a pointer in a DLL, so that it can be referenced with the
676 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
677 name is formed by combining <code>__imp_</code> and the function or
678 variable name.</dd>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]679</dl>
680
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]681<p>For example, since the "<tt>.LC0</tt>" variable is defined to be internal, if
682 another module defined a "<tt>.LC0</tt>" variable and was linked with this
683 one, one of the two would be renamed, preventing a collision. Since
684 "<tt>main</tt>" and "<tt>puts</tt>" are external (i.e., lacking any linkage
685 declarations), they are accessible outside of the current module.</p>
686
687<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]688 other than <tt>external</tt>, <tt>dllimport</tt>
689 or <tt>extern_weak</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]690
Duncan Sands667d4b82009-03-07 15:45:40 +0000[diff] [blame]691<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]692 or <tt>weak_odr</tt> linkages.</p>
693
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]694</div>
695
696<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]697<h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]698 <a name="callingconv">Calling Conventions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]699</h3>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]700
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]701<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]702
703<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]704 and <a href="#i_invoke">invokes</a> can all have an optional calling
705 convention specified for the call. The calling convention of any pair of
706 dynamic caller/callee must match, or the behavior of the program is
707 undefined. The following calling conventions are supported by LLVM, and more
708 may be added in the future:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]709
710<dl>
711 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]712 <dd>This calling convention (the default if no other calling convention is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]713 specified) matches the target C calling conventions. This calling
714 convention supports varargs function calls and tolerates some mismatch in
715 the declared prototype and implemented declaration of the function (as
716 does normal C).</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]717
718 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]719 <dd>This calling convention attempts to make calls as fast as possible
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]720 (e.g. by passing things in registers). This calling convention allows the
721 target to use whatever tricks it wants to produce fast code for the
722 target, without having to conform to an externally specified ABI
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]723 (Application Binary Interface).
724 <a href="CodeGenerator.html#tailcallopt">Tail calls can only be optimized
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]725 when this or the GHC convention is used.</a> This calling convention
726 does not support varargs and requires the prototype of all callees to
727 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]728
729 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]730 <dd>This calling convention attempts to make code in the caller as efficient
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]731 as possible under the assumption that the call is not commonly executed.
732 As such, these calls often preserve all registers so that the call does
733 not break any live ranges in the caller side. This calling convention
734 does not support varargs and requires the prototype of all callees to
735 exactly match the prototype of the function definition.</dd>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]736
Chris Lattner29689432010-03-11 00:22:57 +0000[diff] [blame]737 <dt><b>"<tt>cc <em>10</em></tt>" - GHC convention</b>:</dt>
738 <dd>This calling convention has been implemented specifically for use by the
739 <a href="http://www.haskell.org/ghc">Glasgow Haskell Compiler (GHC)</a>.
740 It passes everything in registers, going to extremes to achieve this by
741 disabling callee save registers. This calling convention should not be
742 used lightly but only for specific situations such as an alternative to
743 the <em>register pinning</em> performance technique often used when
744 implementing functional programming languages.At the moment only X86
745 supports this convention and it has the following limitations:
746 <ul>
747 <li>On <em>X86-32</em> only supports up to 4 bit type parameters. No
748 floating point types are supported.</li>
749 <li>On <em>X86-64</em> only supports up to 10 bit type parameters and
750 6 floating point parameters.</li>
751 </ul>
752 This calling convention supports
753 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> but
754 requires both the caller and callee are using it.
755 </dd>
756
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]757 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]758 <dd>Any calling convention may be specified by number, allowing
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]759 target-specific calling conventions to be used. Target specific calling
760 conventions start at 64.</dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]761</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]762
763<p>More calling conventions can be added/defined on an as-needed basis, to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]764 support Pascal conventions or any other well-known target-independent
765 convention.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]766
767</div>
768
769<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]770<h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]771 <a name="visibility">Visibility Styles</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]772</h3>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]773
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]774<div>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]775
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]776<p>All Global Variables and Functions have one of the following visibility
777 styles:</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]778
779<dl>
780 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]781 <dd>On targets that use the ELF object file format, default visibility means
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]782 that the declaration is visible to other modules and, in shared libraries,
783 means that the declared entity may be overridden. On Darwin, default
784 visibility means that the declaration is visible to other modules. Default
785 visibility corresponds to "external linkage" in the language.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]786
787 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]788 <dd>Two declarations of an object with hidden visibility refer to the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]789 object if they are in the same shared object. Usually, hidden visibility
790 indicates that the symbol will not be placed into the dynamic symbol
791 table, so no other module (executable or shared library) can reference it
792 directly.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]793
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]794 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]795 <dd>On ELF, protected visibility indicates that the symbol will be placed in
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]796 the dynamic symbol table, but that references within the defining module
797 will bind to the local symbol. That is, the symbol cannot be overridden by
798 another module.</dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]799</dl>
800
801</div>
802
803<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]804<h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]805 <a name="namedtypes">Named Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]806</h3>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]807
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]808<div>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]809
810<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]811 it easier to read the IR and make the IR more condensed (particularly when
812 recursive types are involved). An example of a name specification is:</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]813
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]814<pre class="doc_code">
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]815%mytype = type { %mytype*, i32 }
816</pre>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]817
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]818<p>You may give a name to any <a href="#typesystem">type</a> except
Chris Lattnerdc65f222010-08-17 23:26:04 +0000[diff] [blame]819 "<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]820 is expected with the syntax "%mytype".</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]821
822<p>Note that type names are aliases for the structural type that they indicate,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]823 and that you can therefore specify multiple names for the same type. This
824 often leads to confusing behavior when dumping out a .ll file. Since LLVM IR
825 uses structural typing, the name is not part of the type. When printing out
826 LLVM IR, the printer will pick <em>one name</em> to render all types of a
827 particular shape. This means that if you have code where two different
828 source types end up having the same LLVM type, that the dumper will sometimes
829 print the "wrong" or unexpected type. This is an important design point and
830 isn't going to change.</p>
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]831
832</div>
833
Chris Lattnere7886e42009-01-11 20:53:49 +0000[diff] [blame]834<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]835<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]836 <a name="globalvars">Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]837</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]838
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]839<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]840
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]841<p>Global variables define regions of memory allocated at compilation time
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]842 instead of run-time. Global variables may optionally be initialized, may
843 have an explicit section to be placed in, and may have an optional explicit
Hans Wennborgce718ff2012-06-23 11:37:03 +0000[diff] [blame]844 alignment specified.</p>
845
846<p>A variable may be defined as <tt>thread_local</tt>, which
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]847 means that it will not be shared by threads (each thread will have a
Hans Wennborgce718ff2012-06-23 11:37:03 +0000[diff] [blame]848 separated copy of the variable). Not all targets support thread-local
849 variables. Optionally, a TLS model may be specified:</p>
850
851<dl>
852 <dt><b><tt>localdynamic</tt></b>:</dt>
853 <dd>For variables that are only used within the current shared library.</dd>
854
855 <dt><b><tt>initialexec</tt></b>:</dt>
856 <dd>For variables in modules that will not be loaded dynamically.</dd>
857
858 <dt><b><tt>localexec</tt></b>:</dt>
859 <dd>For variables defined in the executable and only used within it.</dd>
860</dl>
861
862<p>The models correspond to the ELF TLS models; see
863 <a href="http://people.redhat.com/drepper/tls.pdf">ELF
864 Handling For Thread-Local Storage</a> for more information on under which
865 circumstances the different models may be used. The target may choose a
866 different TLS model if the specified model is not supported, or if a better
867 choice of model can be made.</p>
868
869<p>A variable may be defined as a global
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]870 "constant," which indicates that the contents of the variable
871 will <b>never</b> be modified (enabling better optimization, allowing the
872 global data to be placed in the read-only section of an executable, etc).
873 Note that variables that need runtime initialization cannot be marked
874 "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]875
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]876<p>LLVM explicitly allows <em>declarations</em> of global variables to be marked
877 constant, even if the final definition of the global is not. This capability
878 can be used to enable slightly better optimization of the program, but
879 requires the language definition to guarantee that optimizations based on the
880 'constantness' are valid for the translation units that do not include the
881 definition.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]882
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]883<p>As SSA values, global variables define pointer values that are in scope
884 (i.e. they dominate) all basic blocks in the program. Global variables
885 always define a pointer to their "content" type because they describe a
886 region of memory, and all memory objects in LLVM are accessed through
887 pointers.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]888
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]889<p>Global variables can be marked with <tt>unnamed_addr</tt> which indicates
890 that the address is not significant, only the content. Constants marked
Rafael Espindolaa5eaa862011-01-15 08:20:57 +0000[diff] [blame]891 like this can be merged with other constants if they have the same
892 initializer. Note that a constant with significant address <em>can</em>
893 be merged with a <tt>unnamed_addr</tt> constant, the result being a
894 constant whose address is significant.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]895
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]896<p>A global variable may be declared to reside in a target-specific numbered
897 address space. For targets that support them, address spaces may affect how
898 optimizations are performed and/or what target instructions are used to
899 access the variable. The default address space is zero. The address space
900 qualifier must precede any other attributes.</p>
Christopher Lamb284d9922007-12-11 09:31:00 +0000[diff] [blame]901
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]902<p>LLVM allows an explicit section to be specified for globals. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]903 supports it, it will emit globals to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]904
Chris Lattnerce99fa92010-04-28 00:13:42 +0000[diff] [blame]905<p>An explicit alignment may be specified for a global, which must be a power
906 of 2. If not present, or if the alignment is set to zero, the alignment of
907 the global is set by the target to whatever it feels convenient. If an
908 explicit alignment is specified, the global is forced to have exactly that
Chris Lattner2d4b8ee2010-04-28 00:31:12 +0000[diff] [blame]909 alignment. Targets and optimizers are not allowed to over-align the global
910 if the global has an assigned section. In this case, the extra alignment
911 could be observable: for example, code could assume that the globals are
912 densely packed in their section and try to iterate over them as an array,
913 alignment padding would break this iteration.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]914
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]915<p>For example, the following defines a global in a numbered address space with
916 an initializer, section, and alignment:</p>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]917
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]918<pre class="doc_code">
Dan Gohman398873c2009-01-11 00:40:00 +0000[diff] [blame]919@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]920</pre>
921
Hans Wennborgce718ff2012-06-23 11:37:03 +0000[diff] [blame]922<p>The following example defines a thread-local global with
923 the <tt>initialexec</tt> TLS model:</p>
924
925<pre class="doc_code">
926@G = thread_local(initialexec) global i32 0, align 4
927</pre>
928
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]929</div>
930
931
932<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]933<h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]934 <a name="functionstructure">Functions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]935</h3>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]936
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]937<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]938
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]939<p>LLVM function definitions consist of the "<tt>define</tt>" keyword, an
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]940 optional <a href="#linkage">linkage type</a>, an optional
941 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]942 <a href="#callingconv">calling convention</a>,
943 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]944 <a href="#paramattrs">parameter attribute</a> for the return type, a function
945 name, a (possibly empty) argument list (each with optional
946 <a href="#paramattrs">parameter attributes</a>), optional
947 <a href="#fnattrs">function attributes</a>, an optional section, an optional
948 alignment, an optional <a href="#gc">garbage collector name</a>, an opening
949 curly brace, a list of basic blocks, and a closing curly brace.</p>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]950
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]951<p>LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
952 optional <a href="#linkage">linkage type</a>, an optional
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]953 <a href="#visibility">visibility style</a>, an optional
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]954 <a href="#callingconv">calling convention</a>,
955 an optional <tt>unnamed_addr</tt> attribute, a return type, an optional
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]956 <a href="#paramattrs">parameter attribute</a> for the return type, a function
957 name, a possibly empty list of arguments, an optional alignment, and an
958 optional <a href="#gc">garbage collector name</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]959
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]960<p>A function definition contains a list of basic blocks, forming the CFG
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]961 (Control Flow Graph) for the function. Each basic block may optionally start
962 with a label (giving the basic block a symbol table entry), contains a list
963 of instructions, and ends with a <a href="#terminators">terminator</a>
964 instruction (such as a branch or function return).</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]965
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]966<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]967 executed on entrance to the function, and it is not allowed to have
968 predecessor basic blocks (i.e. there can not be any branches to the entry
969 block of a function). Because the block can have no predecessors, it also
970 cannot have any <a href="#i_phi">PHI nodes</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]971
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]972<p>LLVM allows an explicit section to be specified for functions. If the target
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]973 supports it, it will emit functions to the section specified.</p>
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]974
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]975<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]976 the alignment is set to zero, the alignment of the function is set by the
977 target to whatever it feels convenient. If an explicit alignment is
978 specified, the function is forced to have at least that much alignment. All
979 alignments must be a power of 2.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]980
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]981<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]982 be significant and two identical functions can be merged.</p>
Rafael Espindolabea46262011-01-08 16:42:36 +0000[diff] [blame]983
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]984<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]985<pre class="doc_code">
Chris Lattner50ad45c2008-10-13 16:55:18 +0000[diff] [blame]986define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]987 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
988 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
989 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
990 [<a href="#gc">gc</a>] { ... }
991</pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]992
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]993</div>
994
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]995<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]996<h3>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]997 <a name="aliasstructure">Aliases</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]998</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]999
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1000<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1001
1002<p>Aliases act as "second name" for the aliasee value (which can be either
1003 function, global variable, another alias or bitcast of global value). Aliases
1004 may have an optional <a href="#linkage">linkage type</a>, and an
1005 optional <a href="#visibility">visibility style</a>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]1006
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1007<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1008<pre class="doc_code">
Duncan Sands0b23ac12008-09-12 20:48:21 +0000[diff] [blame]1009@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]1010</pre>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]1011
1012</div>
1013
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1014<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1015<h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]1016 <a name="namedmetadatastructure">Named Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1017</h3>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]1018
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1019<div>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]1020
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]1021<p>Named metadata is a collection of metadata. <a href="#metadata">Metadata
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]1022 nodes</a> (but not metadata strings) are the only valid operands for
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]1023 a named metadata.</p>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]1024
1025<h5>Syntax:</h5>
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1026<pre class="doc_code">
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]1027; Some unnamed metadata nodes, which are referenced by the named metadata.
1028!0 = metadata !{metadata !"zero"}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]1029!1 = metadata !{metadata !"one"}
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]1030!2 = metadata !{metadata !"two"}
Dan Gohman1005bc52010-07-13 19:48:13 +0000[diff] [blame]1031; A named metadata.
Dan Gohman872814a2010-07-21 18:54:18 +0000[diff] [blame]1032!name = !{!0, !1, !2}
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]1033</pre>
Devang Patelcd1fd252010-01-11 19:35:55 +0000[diff] [blame]1034
1035</div>
1036
1037<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1038<h3>
1039 <a name="paramattrs">Parameter Attributes</a>
1040</h3>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1041
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1042<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1043
1044<p>The return type and each parameter of a function type may have a set of
1045 <i>parameter attributes</i> associated with them. Parameter attributes are
1046 used to communicate additional information about the result or parameters of
1047 a function. Parameter attributes are considered to be part of the function,
1048 not of the function type, so functions with different parameter attributes
1049 can have the same function type.</p>
1050
1051<p>Parameter attributes are simple keywords that follow the type specified. If
1052 multiple parameter attributes are needed, they are space separated. For
1053 example:</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1054
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1055<pre class="doc_code">
Nick Lewyckyb6a7d252009-02-15 23:06:14 +0000[diff] [blame]1056declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner66d922c2008-10-04 18:33:34 +0000[diff] [blame]1057declare i32 @atoi(i8 zeroext)
1058declare signext i8 @returns_signed_char()
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1059</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1060
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1061<p>Note that any attributes for the function result (<tt>nounwind</tt>,
1062 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1063
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1064<p>Currently, only the following parameter attributes are defined:</p>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1065
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1066<dl>
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1067 <dt><tt><b>zeroext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1068 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarichebe81732011-03-16 22:20:18 +0000[diff] [blame]1069 should be zero-extended to the extent required by the target's ABI (which
1070 is usually 32-bits, but is 8-bits for a i1 on x86-64) by the caller (for a
1071 parameter) or the callee (for a return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1072
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1073 <dt><tt><b>signext</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1074 <dd>This indicates to the code generator that the parameter or return value
Cameron Zwarich9e69ff92011-03-17 14:21:58 +0000[diff] [blame]1075 should be sign-extended to the extent required by the target's ABI (which
1076 is usually 32-bits) by the caller (for a parameter) or the callee (for a
1077 return value).</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1078
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1079 <dt><tt><b>inreg</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1080 <dd>This indicates that this parameter or return value should be treated in a
1081 special target-dependent fashion during while emitting code for a function
1082 call or return (usually, by putting it in a register as opposed to memory,
1083 though some targets use it to distinguish between two different kinds of
1084 registers). Use of this attribute is target-specific.</dd>
Chris Lattner47507de2008-01-11 06:20:47 +0000[diff] [blame]1085
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1086 <dt><tt><b><a name="byval">byval</a></b></tt></dt>
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1087 <dd><p>This indicates that the pointer parameter should really be passed by
1088 value to the function. The attribute implies that a hidden copy of the
1089 pointee
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1090 is made between the caller and the callee, so the callee is unable to
Chris Lattneref097052012-05-30 00:40:23 +0000[diff] [blame]1091 modify the value in the caller. This attribute is only valid on LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1092 pointer arguments. It is generally used to pass structs and arrays by
1093 value, but is also valid on pointers to scalars. The copy is considered
1094 to belong to the caller not the callee (for example,
1095 <tt><a href="#readonly">readonly</a></tt> functions should not write to
1096 <tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattnera6fd81d2010-11-20 23:49:06 +0000[diff] [blame]1097 values.</p>
1098
1099 <p>The byval attribute also supports specifying an alignment with
1100 the align attribute. It indicates the alignment of the stack slot to
1101 form and the known alignment of the pointer specified to the call site. If
1102 the alignment is not specified, then the code generator makes a
1103 target-specific assumption.</p></dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1104
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1105 <dt><tt><b><a name="sret">sret</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1106 <dd>This indicates that the pointer parameter specifies the address of a
1107 structure that is the return value of the function in the source program.
1108 This pointer must be guaranteed by the caller to be valid: loads and
1109 stores to the structure may be assumed by the callee to not to trap. This
1110 may only be applied to the first parameter. This is not a valid attribute
1111 for return values. </dd>
1112
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1113 <dt><tt><b><a name="noalias">noalias</a></b></tt></dt>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1114 <dd>This indicates that pointer values
1115 <a href="#pointeraliasing"><i>based</i></a> on the argument or return
Dan Gohmanefca7f92010-07-02 23:46:54 +0000[diff] [blame]1116 value do not alias pointer values which are not <i>based</i> on it,
1117 ignoring certain "irrelevant" dependencies.
1118 For a call to the parent function, dependencies between memory
1119 references from before or after the call and from those during the call
1120 are "irrelevant" to the <tt>noalias</tt> keyword for the arguments and
1121 return value used in that call.
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1122 The caller shares the responsibility with the callee for ensuring that
1123 these requirements are met.
1124 For further details, please see the discussion of the NoAlias response in
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1125 <a href="AliasAnalysis.html#MustMayNo">alias analysis</a>.<br>
1126<br>
John McCall191d4ee2010-07-06 21:07:14 +0000[diff] [blame]1127 Note that this definition of <tt>noalias</tt> is intentionally
1128 similar to the definition of <tt>restrict</tt> in C99 for function
Chris Lattner211244a2010-07-06 20:51:35 +0000[diff] [blame]1129 arguments, though it is slightly weaker.
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]1130<br>
1131 For function return values, C99's <tt>restrict</tt> is not meaningful,
1132 while LLVM's <tt>noalias</tt> is.
1133 </dd>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1134
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1135 <dt><tt><b><a name="nocapture">nocapture</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1136 <dd>This indicates that the callee does not make any copies of the pointer
1137 that outlive the callee itself. This is not a valid attribute for return
1138 values.</dd>
1139
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]1140 <dt><tt><b><a name="nest">nest</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1141 <dd>This indicates that the pointer parameter can be excised using the
1142 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
1143 attribute for return values.</dd>
1144</dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1145
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1146</div>
1147
1148<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1149<h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1150 <a name="gc">Garbage Collector Names</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1151</h3>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1152
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1153<div>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1154
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1155<p>Each function may specify a garbage collector name, which is simply a
1156 string:</p>
1157
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1158<pre class="doc_code">
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1159define void @f() gc "name" { ... }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1160</pre>
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1161
1162<p>The compiler declares the supported values of <i>name</i>. Specifying a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1163 collector which will cause the compiler to alter its output in order to
1164 support the named garbage collection algorithm.</p>
1165
Gordon Henriksen80a75bf2007-12-10 03:18:06 +0000[diff] [blame]1166</div>
1167
1168<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1169<h3>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1170 <a name="fnattrs">Function Attributes</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1171</h3>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1172
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1173<div>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1174
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1175<p>Function attributes are set to communicate additional information about a
1176 function. Function attributes are considered to be part of the function, not
1177 of the function type, so functions with different parameter attributes can
1178 have the same function type.</p>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1179
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1180<p>Function attributes are simple keywords that follow the type specified. If
1181 multiple attributes are needed, they are space separated. For example:</p>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1182
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1183<pre class="doc_code">
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1184define void @f() noinline { ... }
1185define void @f() alwaysinline { ... }
1186define void @f() alwaysinline optsize { ... }
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1187define void @f() optsize { ... }
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1188</pre>
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1189
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1190<dl>
Kostya Serebryany164b86b2012-01-20 17:56:17 +0000[diff] [blame]1191 <dt><tt><b>address_safety</b></tt></dt>
1192 <dd>This attribute indicates that the address safety analysis
1193 is enabled for this function. </dd>
1194
Charles Davis1e063d12010-02-12 00:31:15 +0000[diff] [blame]1195 <dt><tt><b>alignstack(&lt;<em>n</em>&gt;)</b></tt></dt>
1196 <dd>This attribute indicates that, when emitting the prologue and epilogue,
1197 the backend should forcibly align the stack pointer. Specify the
1198 desired alignment, which must be a power of two, in parentheses.
1199
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1200 <dt><tt><b>alwaysinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1201 <dd>This attribute indicates that the inliner should attempt to inline this
1202 function into callers whenever possible, ignoring any active inlining size
1203 threshold for this caller.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1204
Dan Gohman129bd562011-06-16 16:03:13 +0000[diff] [blame]1205 <dt><tt><b>nonlazybind</b></tt></dt>
1206 <dd>This attribute suppresses lazy symbol binding for the function. This
1207 may make calls to the function faster, at the cost of extra program
1208 startup time if the function is not called during program startup.</dd>
1209
Jakob Stoklund Olesen570a4a52010-02-06 01:16:28 +0000[diff] [blame]1210 <dt><tt><b>inlinehint</b></tt></dt>
1211 <dd>This attribute indicates that the source code contained a hint that inlining
1212 this function is desirable (such as the "inline" keyword in C/C++). It
1213 is just a hint; it imposes no requirements on the inliner.</dd>
1214
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1215 <dt><tt><b>naked</b></tt></dt>
1216 <dd>This attribute disables prologue / epilogue emission for the function.
1217 This can have very system-specific consequences.</dd>
1218
1219 <dt><tt><b>noimplicitfloat</b></tt></dt>
1220 <dd>This attributes disables implicit floating point instructions.</dd>
1221
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1222 <dt><tt><b>noinline</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1223 <dd>This attribute indicates that the inliner should never inline this
1224 function in any situation. This attribute may not be used together with
1225 the <tt>alwaysinline</tt> attribute.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1226
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1227 <dt><tt><b>noredzone</b></tt></dt>
1228 <dd>This attribute indicates that the code generator should not use a red
1229 zone, even if the target-specific ABI normally permits it.</dd>
Devang Patel2c9c3e72008-09-26 23:51:19 +0000[diff] [blame]1230
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1231 <dt><tt><b>noreturn</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1232 <dd>This function attribute indicates that the function never returns
1233 normally. This produces undefined behavior at runtime if the function
1234 ever does dynamically return.</dd>
Bill Wendling31359ba2008-11-13 01:02:51 +0000[diff] [blame]1235
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1236 <dt><tt><b>nounwind</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1237 <dd>This function attribute indicates that the function never returns with an
1238 unwind or exceptional control flow. If the function does unwind, its
1239 runtime behavior is undefined.</dd>
Bill Wendlingfbaa7ed2008-11-26 19:07:40 +0000[diff] [blame]1240
Nick Lewycky76ec37a2010-07-06 18:24:09 +0000[diff] [blame]1241 <dt><tt><b>optsize</b></tt></dt>
1242 <dd>This attribute suggests that optimization passes and code generator passes
1243 make choices that keep the code size of this function low, and otherwise
1244 do optimizations specifically to reduce code size.</dd>
1245
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1246 <dt><tt><b>readnone</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1247 <dd>This attribute indicates that the function computes its result (or decides
1248 to unwind an exception) based strictly on its arguments, without
1249 dereferencing any pointer arguments or otherwise accessing any mutable
1250 state (e.g. memory, control registers, etc) visible to caller functions.
1251 It does not write through any pointer arguments
1252 (including <tt><a href="#byval">byval</a></tt> arguments) and never
1253 changes any state visible to callers. This means that it cannot unwind
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]1254 exceptions by calling the <tt>C++</tt> exception throwing methods.</dd>
Devang Patel5d96fda2009-06-12 19:45:19 +0000[diff] [blame]1255
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1256 <dt><tt><b><a name="readonly">readonly</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1257 <dd>This attribute indicates that the function does not write through any
1258 pointer arguments (including <tt><a href="#byval">byval</a></tt>
1259 arguments) or otherwise modify any state (e.g. memory, control registers,
1260 etc) visible to caller functions. It may dereference pointer arguments
1261 and read state that may be set in the caller. A readonly function always
1262 returns the same value (or unwinds an exception identically) when called
1263 with the same set of arguments and global state. It cannot unwind an
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]1264 exception by calling the <tt>C++</tt> exception throwing methods.</dd>
Anton Korobeynikovc5ec8a72009-07-17 18:07:26 +0000[diff] [blame]1265
Bill Wendling9bd5d042011-12-05 21:27:54 +0000[diff] [blame]1266 <dt><tt><b><a name="returns_twice">returns_twice</a></b></tt></dt>
1267 <dd>This attribute indicates that this function can return twice. The
1268 C <code>setjmp</code> is an example of such a function. The compiler
1269 disables some optimizations (like tail calls) in the caller of these
1270 functions.</dd>
1271
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1272 <dt><tt><b><a name="ssp">ssp</a></b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1273 <dd>This attribute indicates that the function should emit a stack smashing
1274 protector. It is in the form of a "canary"&mdash;a random value placed on
1275 the stack before the local variables that's checked upon return from the
1276 function to see if it has been overwritten. A heuristic is used to
1277 determine if a function needs stack protectors or not.<br>
1278<br>
1279 If a function that has an <tt>ssp</tt> attribute is inlined into a
1280 function that doesn't have an <tt>ssp</tt> attribute, then the resulting
1281 function will have an <tt>ssp</tt> attribute.</dd>
1282
Bill Wendlingf82d40a2009-11-02 00:24:16 +0000[diff] [blame]1283 <dt><tt><b>sspreq</b></tt></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1284 <dd>This attribute indicates that the function should <em>always</em> emit a
1285 stack smashing protector. This overrides
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1286 the <tt><a href="#ssp">ssp</a></tt> function attribute.<br>
1287<br>
1288 If a function that has an <tt>sspreq</tt> attribute is inlined into a
1289 function that doesn't have an <tt>sspreq</tt> attribute or which has
1290 an <tt>ssp</tt> attribute, then the resulting function will have
1291 an <tt>sspreq</tt> attribute.</dd>
Rafael Espindolafbff0ec2011-07-25 15:27:59 +0000[diff] [blame]1292
1293 <dt><tt><b><a name="uwtable">uwtable</a></b></tt></dt>
1294 <dd>This attribute indicates that the ABI being targeted requires that
1295 an unwind table entry be produce for this function even if we can
1296 show that no exceptions passes by it. This is normally the case for
1297 the ELF x86-64 abi, but it can be disabled for some compilation
1298 units.</dd>
Bill Wendlinge36dccc2008-09-07 10:26:33 +0000[diff] [blame]1299</dl>
1300
Devang Patelf8b94812008-09-04 23:05:13 +0000[diff] [blame]1301</div>
1302
1303<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1304<h3>
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]1305 <a name="moduleasm">Module-Level Inline Assembly</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1306</h3>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1307
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1308<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1309
1310<p>Modules may contain "module-level inline asm" blocks, which corresponds to
1311 the GCC "file scope inline asm" blocks. These blocks are internally
1312 concatenated by LLVM and treated as a single unit, but may be separated in
1313 the <tt>.ll</tt> file if desired. The syntax is very simple:</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1314
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1315<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1316module asm "inline asm code goes here"
1317module asm "more can go here"
1318</pre>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1319
1320<p>The strings can contain any character by escaping non-printable characters.
1321 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]1322 for the number.</p>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1323
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1324<p>The inline asm code is simply printed to the machine code .s file when
1325 assembly code is generated.</p>
1326
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]1327</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1328
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1329<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1330<h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1331 <a name="datalayout">Data Layout</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1332</h3>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1333
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1334<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1335
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1336<p>A module may specify a target specific data layout string that specifies how
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1337 data is to be laid out in memory. The syntax for the data layout is
1338 simply:</p>
1339
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]1340<pre class="doc_code">
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1341target datalayout = "<i>layout specification</i>"
1342</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1343
1344<p>The <i>layout specification</i> consists of a list of specifications
1345 separated by the minus sign character ('-'). Each specification starts with
1346 a letter and may include other information after the letter to define some
1347 aspect of the data layout. The specifications accepted are as follows:</p>
1348
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1349<dl>
1350 <dt><tt>E</tt></dt>
1351 <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]1352 bits with the most significance have the lowest address location.</dd>
1353
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1354 <dt><tt>e</tt></dt>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1355 <dd>Specifies that the target lays out data in little-endian form. That is,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1356 the bits with the least significance have the lowest address
1357 location.</dd>
1358
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1359 <dt><tt>S<i>size</i></tt></dt>
1360 <dd>Specifies the natural alignment of the stack in bits. Alignment promotion
1361 of stack variables is limited to the natural stack alignment to avoid
1362 dynamic stack realignment. The stack alignment must be a multiple of
Lang Hames5f119a62011-10-11 17:50:14 +0000[diff] [blame]1363 8-bits. If omitted, the natural stack alignment defaults to "unspecified",
1364 which does not prevent any alignment promotions.</dd>
Lang Hamesbb5b3f32011-10-10 23:42:08 +0000[diff] [blame]1365
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1366 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1367 <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]1368 <i>preferred</i> alignments. All sizes are in bits. Specifying
1369 the <i>pref</i> alignment is optional. If omitted, the
1370 preceding <tt>:</tt> should be omitted too.</dd>
1371
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1372 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1373 <dd>This specifies the alignment for an integer type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1374 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1375
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1376 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1377 <dd>This specifies the alignment for a vector type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1378 <i>size</i>.</dd>
1379
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1380 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]1381 <dd>This specifies the alignment for a floating point type of a given bit
Dale Johannesen9d8d2212010-05-28 18:54:47 +0000[diff] [blame]1382 <i>size</i>. Only values of <i>size</i> that are supported by the target
1383 will work. 32 (float) and 64 (double) are supported on all targets;
1384 80 or 128 (different flavors of long double) are also supported on some
1385 targets.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1386
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1387 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1388 <dd>This specifies the alignment for an aggregate type of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1389 <i>size</i>.</dd>
1390
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1391 <dt><tt>s<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1392 <dd>This specifies the alignment for a stack object of a given bit
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1393 <i>size</i>.</dd>
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1394
1395 <dt><tt>n<i>size1</i>:<i>size2</i>:<i>size3</i>...</tt></dt>
1396 <dd>This specifies a set of native integer widths for the target CPU
1397 in bits. For example, it might contain "n32" for 32-bit PowerPC,
1398 "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]1399 this set are considered to support most general arithmetic
Chris Lattnere82bdc42009-11-07 09:35:34 +0000[diff] [blame]1400 operations efficiently.</dd>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1401</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1402
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1403<p>When constructing the data layout for a given target, LLVM starts with a
Dan Gohman1c70c002010-04-28 00:36:01 +0000[diff] [blame]1404 default set of specifications which are then (possibly) overridden by the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1405 specifications in the <tt>datalayout</tt> keyword. The default specifications
1406 are given in this list:</p>
1407
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1408<ul>
1409 <li><tt>E</tt> - big endian</li>
Dan Gohmanfdf2e8c2010-02-23 02:44:03 +0000[diff] [blame]1410 <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]1411 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1412 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1413 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1414 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]1415 <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]1416 alignment of 64-bits</li>
1417 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1418 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1419 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1420 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1421 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
Daniel Dunbar87bde0b2009-06-08 22:17:53 +0000[diff] [blame]1422 <li><tt>s0:64:64</tt> - stack objects are 64-bit aligned</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1423</ul>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1424
1425<p>When LLVM is determining the alignment for a given type, it uses the
1426 following rules:</p>
1427
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1428<ol>
1429 <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]1430 specification is used.</li>
1431
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1432 <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]1433 smallest integer type that is larger than the bitwidth of the sought type
1434 is used. If none of the specifications are larger than the bitwidth then
Sylvestre Ledruc8e41c52012-07-23 08:51:15 +0000[diff] [blame]1435 the largest integer type is used. For example, given the default
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1436 specifications above, the i7 type will use the alignment of i8 (next
1437 largest) while both i65 and i256 will use the alignment of i64 (largest
1438 specified).</li>
1439
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1440 <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]1441 largest vector type that is smaller than the sought vector type will be
1442 used as a fall back. This happens because &lt;128 x double&gt; can be
1443 implemented in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1444</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1445
Chris Lattner6509f502011-10-11 23:01:39 +0000[diff] [blame]1446<p>The function of the data layout string may not be what you expect. Notably,
1447 this is not a specification from the frontend of what alignment the code
1448 generator should use.</p>
1449
1450<p>Instead, if specified, the target data layout is required to match what the
1451 ultimate <em>code generator</em> expects. This string is used by the
1452 mid-level optimizers to
1453 improve code, and this only works if it matches what the ultimate code
1454 generator uses. If you would like to generate IR that does not embed this
1455 target-specific detail into the IR, then you don't have to specify the
1456 string. This will disable some optimizations that require precise layout
1457 information, but this also prevents those optimizations from introducing
1458 target specificity into the IR.</p>
1459
1460
1461
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1462</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1463
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1464<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1465<h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1466 <a name="pointeraliasing">Pointer Aliasing Rules</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1467</h3>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1468
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1469<div>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1470
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1471<p>Any memory access must be done through a pointer value associated
Andreas Bolka99a82052009-07-27 20:37:10 +0000[diff] [blame]1472with an address range of the memory access, otherwise the behavior
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1473is undefined. Pointer values are associated with address ranges
1474according to the following rules:</p>
1475
1476<ul>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1477 <li>A pointer value is associated with the addresses associated with
1478 any value it is <i>based</i> on.
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1479 <li>An address of a global variable is associated with the address
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1480 range of the variable's storage.</li>
1481 <li>The result value of an allocation instruction is associated with
1482 the address range of the allocated storage.</li>
1483 <li>A null pointer in the default address-space is associated with
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1484 no address.</li>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1485 <li>An integer constant other than zero or a pointer value returned
1486 from a function not defined within LLVM may be associated with address
1487 ranges allocated through mechanisms other than those provided by
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1488 LLVM. Such ranges shall not overlap with any ranges of addresses
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1489 allocated by mechanisms provided by LLVM.</li>
Dan Gohman1e109622010-07-02 18:41:32 +0000[diff] [blame]1490</ul>
1491
1492<p>A pointer value is <i>based</i> on another pointer value according
1493 to the following rules:</p>
1494
1495<ul>
1496 <li>A pointer value formed from a
1497 <tt><a href="#i_getelementptr">getelementptr</a></tt> operation
1498 is <i>based</i> on the first operand of the <tt>getelementptr</tt>.</li>
1499 <li>The result value of a
1500 <tt><a href="#i_bitcast">bitcast</a></tt> is <i>based</i> on the operand
1501 of the <tt>bitcast</tt>.</li>
1502 <li>A pointer value formed by an
1503 <tt><a href="#i_inttoptr">inttoptr</a></tt> is <i>based</i> on all
1504 pointer values that contribute (directly or indirectly) to the
1505 computation of the pointer's value.</li>
1506 <li>The "<i>based</i> on" relationship is transitive.</li>
1507</ul>
1508
1509<p>Note that this definition of <i>"based"</i> is intentionally
1510 similar to the definition of <i>"based"</i> in C99, though it is
1511 slightly weaker.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1512
1513<p>LLVM IR does not associate types with memory. The result type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1514<tt><a href="#i_load">load</a></tt> merely indicates the size and
1515alignment of the memory from which to load, as well as the
Dan Gohmanc22c0f32010-06-17 19:23:50 +0000[diff] [blame]1516interpretation of the value. The first operand type of a
Andreas Bolka55e459a2009-07-29 00:02:05 +0000[diff] [blame]1517<tt><a href="#i_store">store</a></tt> similarly only indicates the size
1518and alignment of the store.</p>
Dan Gohman556ca272009-07-27 18:07:55 +0000[diff] [blame]1519
1520<p>Consequently, type-based alias analysis, aka TBAA, aka
1521<tt>-fstrict-aliasing</tt>, is not applicable to general unadorned
1522LLVM IR. <a href="#metadata">Metadata</a> may be used to encode
1523additional information which specialized optimization passes may use
1524to implement type-based alias analysis.</p>
1525
1526</div>
1527
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1528<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1529<h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1530 <a name="volatile">Volatile Memory Accesses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1531</h3>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1532
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1533<div>
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]1534
1535<p>Certain memory accesses, such as <a href="#i_load"><tt>load</tt></a>s, <a
1536href="#i_store"><tt>store</tt></a>s, and <a
1537href="#int_memcpy"><tt>llvm.memcpy</tt></a>s may be marked <tt>volatile</tt>.
1538The optimizers must not change the number of volatile operations or change their
1539order of execution relative to other volatile operations. The optimizers
1540<i>may</i> change the order of volatile operations relative to non-volatile
1541operations. This is not Java's "volatile" and has no cross-thread
1542synchronization behavior.</p>
1543
1544</div>
1545
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1546<!-- ======================================================================= -->
1547<h3>
1548 <a name="memmodel">Memory Model for Concurrent Operations</a>
1549</h3>
1550
1551<div>
1552
1553<p>The LLVM IR does not define any way to start parallel threads of execution
1554or to register signal handlers. Nonetheless, there are platform-specific
1555ways to create them, and we define LLVM IR's behavior in their presence. This
1556model is inspired by the C++0x memory model.</p>
1557
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1558<p>For a more informal introduction to this model, see the
1559<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.
1560
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1561<p>We define a <i>happens-before</i> partial order as the least partial order
1562that</p>
1563<ul>
1564 <li>Is a superset of single-thread program order, and</li>
1565 <li>When a <i>synchronizes-with</i> <tt>b</tt>, includes an edge from
1566 <tt>a</tt> to <tt>b</tt>. <i>Synchronizes-with</i> pairs are introduced
1567 by platform-specific techniques, like pthread locks, thread
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1568 creation, thread joining, etc., and by atomic instructions.
1569 (See also <a href="#ordering">Atomic Memory Ordering Constraints</a>).
1570 </li>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1571</ul>
1572
1573<p>Note that program order does not introduce <i>happens-before</i> edges
1574between a thread and signals executing inside that thread.</p>
1575
1576<p>Every (defined) read operation (load instructions, memcpy, atomic
1577loads/read-modify-writes, etc.) <var>R</var> reads a series of bytes written by
1578(defined) write operations (store instructions, atomic
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1579stores/read-modify-writes, memcpy, etc.). For the purposes of this section,
1580initialized globals are considered to have a write of the initializer which is
1581atomic and happens before any other read or write of the memory in question.
1582For each byte of a read <var>R</var>, <var>R<sub>byte</sub></var> may see
1583any write to the same byte, except:</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1584
1585<ul>
1586 <li>If <var>write<sub>1</sub></var> happens before
1587 <var>write<sub>2</sub></var>, and <var>write<sub>2</sub></var> happens
1588 before <var>R<sub>byte</sub></var>, then <var>R<sub>byte</sub></var>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1589 does not see <var>write<sub>1</sub></var>.
Bill Wendling0246bb72011-07-31 06:45:03 +0000[diff] [blame]1590 <li>If <var>R<sub>byte</sub></var> happens before
1591 <var>write<sub>3</sub></var>, then <var>R<sub>byte</sub></var> does not
1592 see <var>write<sub>3</sub></var>.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1593</ul>
1594
1595<p>Given that definition, <var>R<sub>byte</sub></var> is defined as follows:
1596<ul>
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1597 <li>If <var>R</var> is volatile, the result is target-dependent. (Volatile
1598 is supposed to give guarantees which can support
1599 <code>sig_atomic_t</code> in C/C++, and may be used for accesses to
1600 addresses which do not behave like normal memory. It does not generally
1601 provide cross-thread synchronization.)
1602 <li>Otherwise, if there is no write to the same byte that happens before
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1603 <var>R<sub>byte</sub></var>, <var>R<sub>byte</sub></var> returns
1604 <tt>undef</tt> for that byte.
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1605 <li>Otherwise, if <var>R<sub>byte</sub></var> may see exactly one write,
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1606 <var>R<sub>byte</sub></var> returns the value written by that
1607 write.</li>
Eli Friedman118973a2011-07-22 03:04:45 +0000[diff] [blame]1608 <li>Otherwise, if <var>R</var> is atomic, and all the writes
1609 <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]1610 values written. See the <a href="#ordering">Atomic Memory Ordering
1611 Constraints</a> section for additional constraints on how the choice
1612 is made.
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1613 <li>Otherwise <var>R<sub>byte</sub></var> returns <tt>undef</tt>.</li>
1614</ul>
1615
1616<p><var>R</var> returns the value composed of the series of bytes it read.
1617This implies that some bytes within the value may be <tt>undef</tt>
1618<b>without</b> the entire value being <tt>undef</tt>. Note that this only
1619defines the semantics of the operation; it doesn't mean that targets will
1620emit more than one instruction to read the series of bytes.</p>
1621
1622<p>Note that in cases where none of the atomic intrinsics are used, this model
1623places only one restriction on IR transformations on top of what is required
1624for single-threaded execution: introducing a store to a byte which might not
Eli Friedman101c81d2011-08-02 01:15:34 +0000[diff] [blame]1625otherwise be stored is not allowed in general. (Specifically, in the case
1626where another thread might write to and read from an address, introducing a
1627store can change a load that may see exactly one write into a load that may
1628see multiple writes.)</p>
Eli Friedman5b60e1b2011-07-20 21:35:53 +0000[diff] [blame]1629
1630<!-- FIXME: This model assumes all targets where concurrency is relevant have
1631a byte-size store which doesn't affect adjacent bytes. As far as I can tell,
1632none of the backends currently in the tree fall into this category; however,
1633there might be targets which care. If there are, we want a paragraph
1634like the following:
1635
1636Targets may specify that stores narrower than a certain width are not
1637available; on such a target, for the purposes of this model, treat any
1638non-atomic write with an alignment or width less than the minimum width
1639as if it writes to the relevant surrounding bytes.
1640-->
1641
1642</div>
1643
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1644<!-- ======================================================================= -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1645<h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1646 <a name="ordering">Atomic Memory Ordering Constraints</a>
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1647</h3>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1648
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]1649<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1650
1651<p>Atomic instructions (<a href="#i_cmpxchg"><code>cmpxchg</code></a>,
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]1652<a href="#i_atomicrmw"><code>atomicrmw</code></a>,
1653<a href="#i_fence"><code>fence</code></a>,
1654<a href="#i_load"><code>atomic load</code></a>, and
Eli Friedman8fa281a2011-08-09 23:26:12 +0000[diff] [blame]1655<a href="#i_store"><code>atomic store</code></a>) take an ordering parameter
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1656that determines which other atomic instructions on the same address they
1657<i>synchronize with</i>. These semantics are borrowed from Java and C++0x,
1658but are somewhat more colloquial. If these descriptions aren't precise enough,
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1659check those specs (see spec references in the
Nick Lewycky300a2632012-01-23 08:47:21 +0000[diff] [blame]1660<a href="Atomics.html#introduction">atomics guide</a>).
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1661<a href="#i_fence"><code>fence</code></a> instructions
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1662treat these orderings somewhat differently since they don't take an address.
1663See that instruction's documentation for details.</p>
1664
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1665<p>For a simpler introduction to the ordering constraints, see the
1666<a href="Atomics.html">LLVM Atomic Instructions and Concurrency Guide</a>.</p>
1667
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1668<dl>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1669<dt><code>unordered</code></dt>
1670<dd>The set of values that can be read is governed by the happens-before
1671partial order. A value cannot be read unless some operation wrote it.
1672This is intended to provide a guarantee strong enough to model Java's
1673non-volatile shared variables. This ordering cannot be specified for
1674read-modify-write operations; it is not strong enough to make them atomic
1675in any interesting way.</dd>
1676<dt><code>monotonic</code></dt>
1677<dd>In addition to the guarantees of <code>unordered</code>, there is a single
1678total order for modifications by <code>monotonic</code> operations on each
1679address. All modification orders must be compatible with the happens-before
1680order. There is no guarantee that the modification orders can be combined to
1681a global total order for the whole program (and this often will not be
1682possible). The read in an atomic read-modify-write operation
1683(<a href="#i_cmpxchg"><code>cmpxchg</code></a> and
1684<a href="#i_atomicrmw"><code>atomicrmw</code></a>)
1685reads the value in the modification order immediately before the value it
1686writes. If one atomic read happens before another atomic read of the same
1687address, the later read must see the same value or a later value in the
1688address's modification order. This disallows reordering of
1689<code>monotonic</code> (or stronger) operations on the same address. If an
1690address is written <code>monotonic</code>ally by one thread, and other threads
1691<code>monotonic</code>ally read that address repeatedly, the other threads must
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1692eventually see the write. This corresponds to the C++0x/C1x
1693<code>memory_order_relaxed</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1694<dt><code>acquire</code></dt>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1695<dd>In addition to the guarantees of <code>monotonic</code>,
Eli Friedmanc264b2f2011-08-24 20:28:39 +0000[diff] [blame]1696a <i>synchronizes-with</i> edge may be formed with a <code>release</code>
1697operation. This is intended to model C++'s <code>memory_order_acquire</code>.</dd>
1698<dt><code>release</code></dt>
1699<dd>In addition to the guarantees of <code>monotonic</code>, if this operation
1700writes a value which is subsequently read by an <code>acquire</code> operation,
1701it <i>synchronizes-with</i> that operation. (This isn't a complete
1702description; see the C++0x definition of a release sequence.) This corresponds
1703to the C++0x/C1x <code>memory_order_release</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1704<dt><code>acq_rel</code> (acquire+release)</dt><dd>Acts as both an
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1705<code>acquire</code> and <code>release</code> operation on its address.
1706This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1707<dt><code>seq_cst</code> (sequentially consistent)</dt><dd>
1708<dd>In addition to the guarantees of <code>acq_rel</code>
1709(<code>acquire</code> for an operation which only reads, <code>release</code>
1710for an operation which only writes), there is a global total order on all
1711sequentially-consistent operations on all addresses, which is consistent with
1712the <i>happens-before</i> partial order and with the modification orders of
1713all the affected addresses. Each sequentially-consistent read sees the last
Eli Friedman234bccd2011-08-22 21:35:27 +0000[diff] [blame]1714preceding write to the same address in this global order. This corresponds
1715to the C++0x/C1x <code>memory_order_seq_cst</code> and Java volatile.</dd>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]1716</dl>
1717
1718<p id="singlethread">If an atomic operation is marked <code>singlethread</code>,
1719it only <i>synchronizes with</i> or participates in modification and seq_cst
1720total orderings with other operations running in the same thread (for example,
1721in signal handlers).</p>
1722
1723</div>
1724
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1725</div>
1726
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1727<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1728<h2><a name="typesystem">Type System</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1729<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1730
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1731<div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1732
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1733<p>The LLVM type system is one of the most important features of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1734 intermediate representation. Being typed enables a number of optimizations
1735 to be performed on the intermediate representation directly, without having
1736 to do extra analyses on the side before the transformation. A strong type
1737 system makes it easier to read the generated code and enables novel analyses
1738 and transformations that are not feasible to perform on normal three address
1739 code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]1740
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1741<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1742<h3>
1743 <a name="t_classifications">Type Classifications</a>
1744</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1745
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1746<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1747
1748<p>The types fall into a few useful classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1749
1750<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1751 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1752 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1753 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1754 <td><a href="#t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1755 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1756 </tr>
1757 <tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1758 <td><a href="#t_floating">floating point</a></td>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1759 <td><tt>half, float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1760 </tr>
1761 <tr>
1762 <td><a name="t_firstclass">first class</a></td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1763 <td><a href="#t_integer">integer</a>,
1764 <a href="#t_floating">floating point</a>,
1765 <a href="#t_pointer">pointer</a>,
Dan Gohman0066db62008-06-18 18:42:13 +0000[diff] [blame]1766 <a href="#t_vector">vector</a>,
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]1767 <a href="#t_struct">structure</a>,
1768 <a href="#t_array">array</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1769 <a href="#t_label">label</a>,
1770 <a href="#t_metadata">metadata</a>.
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1771 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1772 </tr>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1773 <tr>
1774 <td><a href="#t_primitive">primitive</a></td>
1775 <td><a href="#t_label">label</a>,
1776 <a href="#t_void">void</a>,
Tobias Grosser05387292010-12-28 20:29:31 +0000[diff] [blame]1777 <a href="#t_integer">integer</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1778 <a href="#t_floating">floating point</a>,
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1779 <a href="#t_x86mmx">x86mmx</a>,
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1780 <a href="#t_metadata">metadata</a>.</td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1781 </tr>
1782 <tr>
1783 <td><a href="#t_derived">derived</a></td>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1784 <td><a href="#t_array">array</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1785 <a href="#t_function">function</a>,
1786 <a href="#t_pointer">pointer</a>,
1787 <a href="#t_struct">structure</a>,
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1788 <a href="#t_vector">vector</a>,
1789 <a href="#t_opaque">opaque</a>.
Dan Gohman01ac1012008-10-14 16:32:04 +0000[diff] [blame]1790 </td>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1791 </tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1792 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1793</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1794
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1795<p>The <a href="#t_firstclass">first class</a> types are perhaps the most
1796 important. Values of these types are the only ones which can be produced by
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1797 instructions.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1798
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1799</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1800
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1801<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1802<h3>
1803 <a name="t_primitive">Primitive Types</a>
1804</h3>
Chris Lattner8f8c7b72008-01-04 04:34:14 +0000[diff] [blame]1805
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1806<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1807
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1808<p>The primitive types are the fundamental building blocks of the LLVM
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1809 system.</p>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1810
1811<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1812<h4>
1813 <a name="t_integer">Integer Type</a>
1814</h4>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1815
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1816<div>
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1817
1818<h5>Overview:</h5>
1819<p>The integer type is a very simple type that simply specifies an arbitrary
1820 bit width for the integer type desired. Any bit width from 1 bit to
1821 2<sup>23</sup>-1 (about 8 million) can be specified.</p>
1822
1823<h5>Syntax:</h5>
1824<pre>
1825 iN
1826</pre>
1827
1828<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1829 value.</p>
1830
1831<h5>Examples:</h5>
1832<table class="layout">
1833 <tr class="layout">
1834 <td class="left"><tt>i1</tt></td>
1835 <td class="left">a single-bit integer.</td>
1836 </tr>
1837 <tr class="layout">
1838 <td class="left"><tt>i32</tt></td>
1839 <td class="left">a 32-bit integer.</td>
1840 </tr>
1841 <tr class="layout">
1842 <td class="left"><tt>i1942652</tt></td>
1843 <td class="left">a really big integer of over 1 million bits.</td>
1844 </tr>
1845</table>
1846
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1847</div>
1848
1849<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1850<h4>
1851 <a name="t_floating">Floating Point Types</a>
1852</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1853
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1854<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1855
1856<table>
1857 <tbody>
1858 <tr><th>Type</th><th>Description</th></tr>
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]1859 <tr><td><tt>half</tt></td><td>16-bit floating point value</td></tr>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1860 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1861 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1862 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1863 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1864 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1865 </tbody>
1866</table>
1867
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1868</div>
1869
1870<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1871<h4>
1872 <a name="t_x86mmx">X86mmx Type</a>
1873</h4>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1874
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1875<div>
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1876
1877<h5>Overview:</h5>
1878<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>
1879
1880<h5>Syntax:</h5>
1881<pre>
Dale Johannesen473a8c82010-10-01 01:07:02 +0000[diff] [blame]1882 x86mmx
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]1883</pre>
1884
1885</div>
1886
1887<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1888<h4>
1889 <a name="t_void">Void Type</a>
1890</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1891
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1892<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1893
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1894<h5>Overview:</h5>
1895<p>The void type does not represent any value and has no size.</p>
1896
1897<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1898<pre>
1899 void
1900</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1901
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1902</div>
1903
1904<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1905<h4>
1906 <a name="t_label">Label Type</a>
1907</h4>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1908
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1909<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1910
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1911<h5>Overview:</h5>
1912<p>The label type represents code labels.</p>
1913
1914<h5>Syntax:</h5>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1915<pre>
1916 label
1917</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1918
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1919</div>
1920
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1921<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1922<h4>
1923 <a name="t_metadata">Metadata Type</a>
1924</h4>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1925
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1926<div>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1927
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1928<h5>Overview:</h5>
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]1929<p>The metadata type represents embedded metadata. No derived types may be
1930 created from metadata except for <a href="#t_function">function</a>
1931 arguments.
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1932
1933<h5>Syntax:</h5>
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1934<pre>
1935 metadata
1936</pre>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]1937
Nick Lewycky7a0370f2009-05-30 05:06:04 +0000[diff] [blame]1938</div>
1939
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1940</div>
Chris Lattner4f69f462008-01-04 04:32:38 +0000[diff] [blame]1941
1942<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1943<h3>
1944 <a name="t_derived">Derived Types</a>
1945</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1946
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1947<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1948
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1949<p>The real power in LLVM comes from the derived types in the system. This is
1950 what allows a programmer to represent arrays, functions, pointers, and other
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]1951 useful types. Each of these types contain one or more element types which
1952 may be a primitive type, or another derived type. For example, it is
1953 possible to have a two dimensional array, using an array as the element type
1954 of another array.</p>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]1955
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1956<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1957<h4>
1958 <a name="t_aggregate">Aggregate Types</a>
1959</h4>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1960
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1961<div>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1962
1963<p>Aggregate Types are a subset of derived types that can contain multiple
Duncan Sands20536b52011-12-14 15:44:20 +0000[diff] [blame]1964 member types. <a href="#t_array">Arrays</a> and
1965 <a href="#t_struct">structs</a> are aggregate types.
1966 <a href="#t_vector">Vectors</a> are not considered to be aggregate types.</p>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]1967
1968</div>
1969
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1970<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]1971<h4>
1972 <a name="t_array">Array Type</a>
1973</h4>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1974
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]1975<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1976
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1977<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1978<p>The array type is a very simple derived type that arranges elements
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1979 sequentially in memory. The array type requires a size (number of elements)
1980 and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1981
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1982<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1983<pre>
1984 [&lt;# elements&gt; x &lt;elementtype&gt;]
1985</pre>
1986
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]1987<p>The number of elements is a constant integer value; <tt>elementtype</tt> may
1988 be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1989
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1990<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1991<table class="layout">
1992 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]1993 <td class="left"><tt>[40 x i32]</tt></td>
1994 <td class="left">Array of 40 32-bit integer values.</td>
1995 </tr>
1996 <tr class="layout">
1997 <td class="left"><tt>[41 x i32]</tt></td>
1998 <td class="left">Array of 41 32-bit integer values.</td>
1999 </tr>
2000 <tr class="layout">
2001 <td class="left"><tt>[4 x i8]</tt></td>
2002 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2003 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2004</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2005<p>Here are some examples of multidimensional arrays:</p>
2006<table class="layout">
2007 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2008 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
2009 <td class="left">3x4 array of 32-bit integer values.</td>
2010 </tr>
2011 <tr class="layout">
2012 <td class="left"><tt>[12 x [10 x float]]</tt></td>
2013 <td class="left">12x10 array of single precision floating point values.</td>
2014 </tr>
2015 <tr class="layout">
2016 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
2017 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2018 </tr>
2019</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2020
Dan Gohman7657f6b2009-11-09 19:01:53 +0000[diff] [blame]2021<p>There is no restriction on indexing beyond the end of the array implied by
2022 a static type (though there are restrictions on indexing beyond the bounds
2023 of an allocated object in some cases). This means that single-dimension
2024 'variable sized array' addressing can be implemented in LLVM with a zero
2025 length array type. An implementation of 'pascal style arrays' in LLVM could
2026 use the type "<tt>{ i32, [0 x float]}</tt>", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2027
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2028</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2029
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2030<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2031<h4>
2032 <a name="t_function">Function Type</a>
2033</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2034
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2035<div>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2036
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2037<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2038<p>The function type can be thought of as a function signature. It consists of
2039 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]2040 function type is a first class type or a void type.</p>
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]2041
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2042<h5>Syntax:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2043<pre>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2044 &lt;returntype&gt; (&lt;parameter list&gt;)
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2045</pre>
2046
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2047<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]2048 specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
2049 which indicates that the function takes a variable number of arguments.
2050 Variable argument functions can access their arguments with
2051 the <a href="#int_varargs">variable argument handling intrinsic</a>
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2052 functions. '<tt>&lt;returntype&gt;</tt>' is any type except
Nick Lewyckyc261df92009-09-27 23:27:42 +0000[diff] [blame]2053 <a href="#t_label">label</a>.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]2054
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2055<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2056<table class="layout">
2057 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2058 <td class="left"><tt>i32 (i32)</tt></td>
2059 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2060 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2061 </tr><tr class="layout">
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2062 <td class="left"><tt>float&nbsp;(i16,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]2063 </tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2064 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]2065 an <tt>i16</tt> and a <a href="#t_pointer">pointer</a> to <tt>i32</tt>,
2066 returning <tt>float</tt>.
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2067 </td>
2068 </tr><tr class="layout">
2069 <td class="left"><tt>i32 (i8*, ...)</tt></td>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2070 <td class="left">A vararg function that takes at least one
2071 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
2072 which returns an integer. This is the signature for <tt>printf</tt> in
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]2073 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2074 </td>
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2075 </tr><tr class="layout">
2076 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Nick Lewycky51386942009-09-27 07:55:32 +0000[diff] [blame]2077 <td class="left">A function taking an <tt>i32</tt>, returning a
2078 <a href="#t_struct">structure</a> containing two <tt>i32</tt> values
Devang Patela582f402008-03-24 05:35:41 +0000[diff] [blame]2079 </td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2080 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2081</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2082
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2083</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2084
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2085<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2086<h4>
2087 <a name="t_struct">Structure Type</a>
2088</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2089
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2090<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2091
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2092<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2093<p>The structure type is used to represent a collection of data members together
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2094 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]2095
Jeffrey Yasskin7a088cf2010-01-11 19:19:26 +0000[diff] [blame]2096<p>Structures in memory are accessed using '<tt><a href="#i_load">load</a></tt>'
2097 and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a field
2098 with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
2099 Structures in registers are accessed using the
2100 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' and
2101 '<tt><a href="#i_insertvalue">insertvalue</a></tt>' instructions.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2102
2103<p>Structures may optionally be "packed" structures, which indicate that the
2104 alignment of the struct is one byte, and that there is no padding between
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2105 the elements. In non-packed structs, padding between field types is inserted
2106 as defined by the TargetData string in the module, which is required to match
Chris Lattnere4617b02011-10-11 23:02:17 +0000[diff] [blame]2107 what the underlying code generator expects.</p>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2108
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2109<p>Structures can either be "literal" or "identified". A literal structure is
2110 defined inline with other types (e.g. <tt>{i32, i32}*</tt>) whereas identified
2111 types are always defined at the top level with a name. Literal types are
2112 uniqued by their contents and can never be recursive or opaque since there is
Chris Lattneraa175c32011-08-12 18:12:40 +0000[diff] [blame]2113 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]2114 never uniqued.
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2115</p>
2116
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2117<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2118<pre>
Chris Lattner2c38d652011-08-12 17:31:02 +0000[diff] [blame]2119 %T1 = type { &lt;type list&gt; } <i>; Identified normal struct type</i>
2120 %T2 = type &lt;{ &lt;type list&gt; }&gt; <i>; Identified packed struct type</i>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2121</pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2122
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2123<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2124<table class="layout">
2125 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2126 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
2127 <td class="left">A triple of three <tt>i32</tt> values</td>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2128 </tr>
2129 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2130 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
2131 <td class="left">A pair, where the first element is a <tt>float</tt> and the
2132 second element is a <a href="#t_pointer">pointer</a> to a
2133 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
2134 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2135 </tr>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2136 <tr class="layout">
2137 <td class="left"><tt>&lt;{ i8, i32 }&gt;</tt></td>
2138 <td class="left">A packed struct known to be 5 bytes in size.</td>
2139 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2140</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2141
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2142</div>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2143
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2144<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2145<h4>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2146 <a name="t_opaque">Opaque Structure Types</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2147</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2148
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2149<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2150
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2151<h5>Overview:</h5>
Chris Lattner628ed392011-07-23 19:59:08 +0000[diff] [blame]2152<p>Opaque structure types are used to represent named structure types that do
2153 not have a body specified. This corresponds (for example) to the C notion of
2154 a forward declared structure.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2155
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2156<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2157<pre>
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2158 %X = type opaque
2159 %52 = type opaque
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2160</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2161
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2162<h5>Examples:</h5>
2163<table class="layout">
2164 <tr class="layout">
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2165 <td class="left"><tt>opaque</tt></td>
2166 <td class="left">An opaque type.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2167 </tr>
2168</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2169
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2170</div>
2171
Chris Lattner1afcace2011-07-09 17:41:24 +0000[diff] [blame]2172
2173
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]2174<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2175<h4>
2176 <a name="t_pointer">Pointer Type</a>
2177</h4>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2178
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2179<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2180
2181<h5>Overview:</h5>
Dan Gohmanff3ef322010-02-25 16:50:07 +0000[diff] [blame]2182<p>The pointer type is used to specify memory locations.
2183 Pointers are commonly used to reference objects in memory.</p>
2184
2185<p>Pointer types may have an optional address space attribute defining the
2186 numbered address space where the pointed-to object resides. The default
2187 address space is number zero. The semantics of non-zero address
2188 spaces are target-specific.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2189
2190<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does it
2191 permit pointers to labels (<tt>label*</tt>). Use <tt>i8*</tt> instead.</p>
Chris Lattner0fd4a272009-02-08 19:53:29 +0000[diff] [blame]2192
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2193<h5>Syntax:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]2194<pre>
2195 &lt;type&gt; *
2196</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2197
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2198<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2199<table class="layout">
2200 <tr class="layout">
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]2201 <td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2202 <td class="left">A <a href="#t_pointer">pointer</a> to <a
2203 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
2204 </tr>
2205 <tr class="layout">
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]2206 <td class="left"><tt>i32 (i32*) *</tt></td>
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2207 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2208 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2209 <tt>i32</tt>.</td>
2210 </tr>
2211 <tr class="layout">
2212 <td class="left"><tt>i32 addrspace(5)*</tt></td>
2213 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
2214 that resides in address space #5.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2215 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2216</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2217
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2218</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2219
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2220<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2221<h4>
2222 <a name="t_vector">Vector Type</a>
2223</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2224
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2225<div>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2226
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2227<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2228<p>A vector type is a simple derived type that represents a vector of elements.
2229 Vector types are used when multiple primitive data are operated in parallel
2230 using a single instruction (SIMD). A vector type requires a size (number of
Duncan Sandsd40d14e2009-11-27 13:38:03 +0000[diff] [blame]2231 elements) and an underlying primitive data type. Vector types are considered
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2232 <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2233
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2234<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2235<pre>
2236 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
2237</pre>
2238
Chris Lattner7d2e7be2010-10-10 18:20:35 +0000[diff] [blame]2239<p>The number of elements is a constant integer value larger than 0; elementtype
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2240 may be any integer or floating point type, or a pointer to these types.
2241 Vectors of size zero are not allowed. </p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]2242
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2243<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2244<table class="layout">
2245 <tr class="layout">
Chris Lattner23ff1f92007-12-19 05:04:11 +0000[diff] [blame]2246 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
2247 <td class="left">Vector of 4 32-bit integer values.</td>
2248 </tr>
2249 <tr class="layout">
2250 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
2251 <td class="left">Vector of 8 32-bit floating-point values.</td>
2252 </tr>
2253 <tr class="layout">
2254 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
2255 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2256 </tr>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]2257 <tr class="layout">
2258 <td class="left"><tt>&lt;4 x i64*&gt;</tt></td>
2259 <td class="left">Vector of 4 pointers to 64-bit integer values.</td>
2260 </tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]2261</table>
Dan Gohmand8791e52009-01-24 15:58:40 +0000[diff] [blame]2262
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2263</div>
2264
Bill Wendlingaf75f0c2011-07-31 06:47:33 +0000[diff] [blame]2265</div>
2266
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]2267</div>
2268
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2269<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2270<h2><a name="constants">Constants</a></h2>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2271<!-- *********************************************************************** -->
2272
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2273<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2274
2275<p>LLVM has several different basic types of constants. This section describes
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2276 them all and their syntax.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2277
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2278<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2279<h3>
2280 <a name="simpleconstants">Simple Constants</a>
2281</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2282
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2283<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2284
2285<dl>
2286 <dt><b>Boolean constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2287 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]2288 constants of the <tt><a href="#t_integer">i1</a></tt> type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2289
2290 <dt><b>Integer constants</b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2291 <dd>Standard integers (such as '4') are constants of
2292 the <a href="#t_integer">integer</a> type. Negative numbers may be used
2293 with integer types.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2294
2295 <dt><b>Floating point constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2296 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2297 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
2298 notation (see below). The assembler requires the exact decimal value of a
2299 floating-point constant. For example, the assembler accepts 1.25 but
2300 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
2301 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2302
2303 <dt><b>Null pointer constants</b></dt>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2304 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2305 and must be of <a href="#t_pointer">pointer type</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2306</dl>
2307
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2308<p>The one non-intuitive notation for constants is the hexadecimal form of
2309 floating point constants. For example, the form '<tt>double
2310 0x432ff973cafa8000</tt>' is equivalent to (but harder to read than)
2311 '<tt>double 4.5e+15</tt>'. The only time hexadecimal floating point
2312 constants are required (and the only time that they are generated by the
2313 disassembler) is when a floating point constant must be emitted but it cannot
2314 be represented as a decimal floating point number in a reasonable number of
2315 digits. For example, NaN's, infinities, and other special values are
2316 represented in their IEEE hexadecimal format so that assembly and disassembly
2317 do not cause any bits to change in the constants.</p>
2318
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2319<p>When using the hexadecimal form, constants of types half, float, and double are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2320 represented using the 16-digit form shown above (which matches the IEEE754
Dan Gohmance163392011-12-17 00:04:22 +0000[diff] [blame]2321 representation for double); half and float values must, however, be exactly
2322 representable as IEE754 half and single precision, respectively.
2323 Hexadecimal format is always used
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2324 for long double, and there are three forms of long double. The 80-bit format
2325 used by x86 is represented as <tt>0xK</tt> followed by 20 hexadecimal digits.
2326 The 128-bit format used by PowerPC (two adjacent doubles) is represented
2327 by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit format
2328 is represented by <tt>0xL</tt> followed by 32 hexadecimal digits; no
2329 currently supported target uses this format. Long doubles will only work if
Tobias Grosser057beb82012-05-24 15:59:06 +0000[diff] [blame]2330 they match the long double format on your target. The IEEE 16-bit format
2331 (half precision) is represented by <tt>0xH</tt> followed by 4 hexadecimal
2332 digits. All hexadecimal formats are big-endian (sign bit at the left).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2333
Dale Johannesen21fe99b2010-10-01 00:48:59 +0000[diff] [blame]2334<p>There are no constants of type x86mmx.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2335</div>
2336
2337<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2338<h3>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]2339<a name="aggregateconstants"></a> <!-- old anchor -->
2340<a name="complexconstants">Complex Constants</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2341</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2342
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2343<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2344
Chris Lattner70882792009-02-28 18:32:25 +0000[diff] [blame]2345<p>Complex constants are a (potentially recursive) combination of simple
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2346 constants and smaller complex constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2347
2348<dl>
2349 <dt><b>Structure constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2350 <dd>Structure constants are represented with notation similar to structure
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2351 type definitions (a comma separated list of elements, surrounded by braces
2352 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
2353 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>".
2354 Structure constants must have <a href="#t_struct">structure type</a>, and
2355 the number and types of elements must match those specified by the
2356 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2357
2358 <dt><b>Array constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2359 <dd>Array constants are represented with notation similar to array type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2360 definitions (a comma separated list of elements, surrounded by square
2361 brackets (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74
2362 ]</tt>". Array constants must have <a href="#t_array">array type</a>, and
2363 the number and types of elements must match those specified by the
2364 type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2365
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2366 <dt><b>Vector constants</b></dt>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2367 <dd>Vector constants are represented with notation similar to vector type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2368 definitions (a comma separated list of elements, surrounded by
2369 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32
2370 42, i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must
2371 have <a href="#t_vector">vector type</a>, and the number and types of
2372 elements must match those specified by the type.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2373
2374 <dt><b>Zero initialization</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2375 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]2376 value to zero of <em>any</em> type, including scalar and
2377 <a href="#t_aggregate">aggregate</a> types.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2378 This is often used to avoid having to print large zero initializers
2379 (e.g. for large arrays) and is always exactly equivalent to using explicit
2380 zero initializers.</dd>
Nick Lewycky21cc4462009-04-04 07:22:01 +0000[diff] [blame]2381
2382 <dt><b>Metadata node</b></dt>
Nick Lewycky1e8c7a62009-05-30 16:08:30 +0000[diff] [blame]2383 <dd>A metadata node is a structure-like constant with
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2384 <a href="#t_metadata">metadata type</a>. For example: "<tt>metadata !{
2385 i32 0, metadata !"test" }</tt>". Unlike other constants that are meant to
2386 be interpreted as part of the instruction stream, metadata is a place to
2387 attach additional information such as debug info.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2388</dl>
2389
2390</div>
2391
2392<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2393<h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2394 <a name="globalconstants">Global Variable and Function Addresses</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2395</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2396
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2397<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2398
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2399<p>The addresses of <a href="#globalvars">global variables</a>
2400 and <a href="#functionstructure">functions</a> are always implicitly valid
2401 (link-time) constants. These constants are explicitly referenced when
2402 the <a href="#identifiers">identifier for the global</a> is used and always
2403 have <a href="#t_pointer">pointer</a> type. For example, the following is a
2404 legal LLVM file:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2405
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2406<pre class="doc_code">
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]2407@X = global i32 17
2408@Y = global i32 42
2409@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2410</pre>
2411
2412</div>
2413
2414<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2415<h3>
2416 <a name="undefvalues">Undefined Values</a>
2417</h3>
2418
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2419<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2420
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2421<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]2422 indicates that the user of the value may receive an unspecified bit-pattern.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2423 Undefined values may be of any type (other than '<tt>label</tt>'
2424 or '<tt>void</tt>') and be used anywhere a constant is permitted.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2425
Chris Lattnerc608cb12009-09-11 01:49:31 +0000[diff] [blame]2426<p>Undefined values are useful because they indicate to the compiler that the
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2427 program is well defined no matter what value is used. This gives the
2428 compiler more freedom to optimize. Here are some examples of (potentially
2429 surprising) transformations that are valid (in pseudo IR):</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2430
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2431
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2432<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2433 %A = add %X, undef
2434 %B = sub %X, undef
2435 %C = xor %X, undef
2436Safe:
2437 %A = undef
2438 %B = undef
2439 %C = undef
2440</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2441
2442<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]2443 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]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 = or %X, undef
2447 %B = and %X, undef
2448Safe:
2449 %A = -1
2450 %B = 0
2451Unsafe:
2452 %A = undef
2453 %B = undef
2454</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2455
2456<p>These logical operations have bits that are not always affected by the input.
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2457 For example, if <tt>%X</tt> has a zero bit, then the output of the
2458 '<tt>and</tt>' operation will always be a zero for that bit, no matter what
2459 the corresponding bit from the '<tt>undef</tt>' is. As such, it is unsafe to
2460 optimize or assume that the result of the '<tt>and</tt>' is '<tt>undef</tt>'.
2461 However, it is safe to assume that all bits of the '<tt>undef</tt>' could be
2462 0, and optimize the '<tt>and</tt>' to 0. Likewise, it is safe to assume that
2463 all the bits of the '<tt>undef</tt>' operand to the '<tt>or</tt>' could be
2464 set, allowing the '<tt>or</tt>' to be folded to -1.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2465
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2466<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2467 %A = select undef, %X, %Y
2468 %B = select undef, 42, %Y
2469 %C = select %X, %Y, undef
2470Safe:
2471 %A = %X (or %Y)
2472 %B = 42 (or %Y)
2473 %C = %Y
2474Unsafe:
2475 %A = undef
2476 %B = undef
2477 %C = undef
2478</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2479
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2480<p>This set of examples shows that undefined '<tt>select</tt>' (and conditional
2481 branch) conditions can go <em>either way</em>, but they have to come from one
2482 of the two operands. In the <tt>%A</tt> example, if <tt>%X</tt> and
2483 <tt>%Y</tt> were both known to have a clear low bit, then <tt>%A</tt> would
2484 have to have a cleared low bit. However, in the <tt>%C</tt> example, the
2485 optimizer is allowed to assume that the '<tt>undef</tt>' operand could be the
2486 same as <tt>%Y</tt>, allowing the whole '<tt>select</tt>' to be
2487 eliminated.</p>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2488
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2489<pre class="doc_code">
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2490 %A = xor undef, undef
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2491
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2492 %B = undef
2493 %C = xor %B, %B
2494
2495 %D = undef
2496 %E = icmp lt %D, 4
2497 %F = icmp gte %D, 4
2498
2499Safe:
2500 %A = undef
2501 %B = undef
2502 %C = undef
2503 %D = undef
2504 %E = undef
2505 %F = undef
2506</pre>
Chris Lattner48a109c2009-09-07 22:52:39 +0000[diff] [blame]2507
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2508<p>This example points out that two '<tt>undef</tt>' operands are not
2509 necessarily the same. This can be surprising to people (and also matches C
2510 semantics) where they assume that "<tt>X^X</tt>" is always zero, even
2511 if <tt>X</tt> is undefined. This isn't true for a number of reasons, but the
2512 short answer is that an '<tt>undef</tt>' "variable" can arbitrarily change
2513 its value over its "live range". This is true because the variable doesn't
2514 actually <em>have a live range</em>. Instead, the value is logically read
2515 from arbitrary registers that happen to be around when needed, so the value
2516 is not necessarily consistent over time. In fact, <tt>%A</tt> and <tt>%C</tt>
2517 need to have the same semantics or the core LLVM "replace all uses with"
2518 concept would not hold.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2519
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2520<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2521 %A = fdiv undef, %X
2522 %B = fdiv %X, undef
2523Safe:
2524 %A = undef
2525b: unreachable
2526</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2527
2528<p>These examples show the crucial difference between an <em>undefined
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2529 value</em> and <em>undefined behavior</em>. An undefined value (like
2530 '<tt>undef</tt>') is allowed to have an arbitrary bit-pattern. This means that
2531 the <tt>%A</tt> operation can be constant folded to '<tt>undef</tt>', because
2532 the '<tt>undef</tt>' could be an SNaN, and <tt>fdiv</tt> is not (currently)
2533 defined on SNaN's. However, in the second example, we can make a more
2534 aggressive assumption: because the <tt>undef</tt> is allowed to be an
2535 arbitrary value, we are allowed to assume that it could be zero. Since a
2536 divide by zero has <em>undefined behavior</em>, we are allowed to assume that
2537 the operation does not execute at all. This allows us to delete the divide and
2538 all code after it. Because the undefined operation "can't happen", the
2539 optimizer can assume that it occurs in dead code.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2540
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2541<pre class="doc_code">
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2542a: store undef -> %X
2543b: store %X -> undef
2544Safe:
2545a: &lt;deleted&gt;
2546b: unreachable
2547</pre>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2548
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2549<p>These examples reiterate the <tt>fdiv</tt> example: a store <em>of</em> an
2550 undefined value can be assumed to not have any effect; we can assume that the
2551 value is overwritten with bits that happen to match what was already there.
2552 However, a store <em>to</em> an undefined location could clobber arbitrary
2553 memory, therefore, it has undefined behavior.</p>
Chris Lattner6e9057b2009-09-07 23:33:52 +0000[diff] [blame]2554
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2555</div>
2556
2557<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2558<h3>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2559 <a name="poisonvalues">Poison Values</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2560</h3>
2561
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2562<div>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2563
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2564<p>Poison values are similar to <a href="#undefvalues">undef values</a>, however
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2565 they also represent the fact that an instruction or constant expression which
2566 cannot evoke side effects has nevertheless detected a condition which results
2567 in undefined behavior.</p>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2568
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2569<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]2570 only exist when produced by operations such as
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2571 <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]2572
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2573<p>Poison value behavior is defined in terms of value <i>dependence</i>:</p>
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2574
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2575<ul>
2576<li>Values other than <a href="#i_phi"><tt>phi</tt></a> nodes depend on
2577 their operands.</li>
2578
2579<li><a href="#i_phi"><tt>Phi</tt></a> nodes depend on the operand corresponding
2580 to their dynamic predecessor basic block.</li>
2581
2582<li>Function arguments depend on the corresponding actual argument values in
2583 the dynamic callers of their functions.</li>
2584
2585<li><a href="#i_call"><tt>Call</tt></a> instructions depend on the
2586 <a href="#i_ret"><tt>ret</tt></a> instructions that dynamically transfer
2587 control back to them.</li>
2588
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2589<li><a href="#i_invoke"><tt>Invoke</tt></a> instructions depend on the
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]2590 <a href="#i_ret"><tt>ret</tt></a>, <a href="#i_resume"><tt>resume</tt></a>,
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2591 or exception-throwing call instructions that dynamically transfer control
2592 back to them.</li>
2593
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2594<li>Non-volatile loads and stores depend on the most recent stores to all of the
2595 referenced memory addresses, following the order in the IR
2596 (including loads and stores implied by intrinsics such as
2597 <a href="#int_memcpy"><tt>@llvm.memcpy</tt></a>.)</li>
2598
Dan Gohman7c24ff12010-05-03 14:59:34 +0000[diff] [blame]2599<!-- TODO: In the case of multiple threads, this only applies if the store
2600 "happens-before" the load or store. -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2601
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2602<!-- TODO: floating-point exception state -->
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2603
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2604<li>An instruction with externally visible side effects depends on the most
2605 recent preceding instruction with externally visible side effects, following
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]2606 the order in the IR. (This includes
2607 <a href="#volatile">volatile operations</a>.)</li>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2608
Dan Gohmanb5328162010-05-03 14:55:22 +0000[diff] [blame]2609<li>An instruction <i>control-depends</i> on a
2610 <a href="#terminators">terminator instruction</a>
2611 if the terminator instruction has multiple successors and the instruction
2612 is always executed when control transfers to one of the successors, and
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2613 may not be executed when control is transferred to another.</li>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2614
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2615<li>Additionally, an instruction also <i>control-depends</i> on a terminator
2616 instruction if the set of instructions it otherwise depends on would be
Chris Lattner7a2bdde2011-04-15 05:18:47 +0000[diff] [blame]2617 different if the terminator had transferred control to a different
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2618 successor.</li>
2619
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2620<li>Dependence is transitive.</li>
2621
2622</ul>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2623
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2624<p>Poison Values have the same behavior as <a href="#undefvalues">undef values</a>,
2625 with the additional affect that any instruction which has a <i>dependence</i>
2626 on a poison value has undefined behavior.</p>
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2627
2628<p>Here are some examples:</p>
Dan Gohmanc30f6e12010-04-26 20:54:53 +0000[diff] [blame]2629
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2630<pre class="doc_code">
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2631entry:
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2632 %poison = sub nuw i32 0, 1 ; Results in a poison value.
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2633 %still_poison = and i32 %poison, 0 ; 0, but also poison.
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2634 %poison_yet_again = getelementptr i32* @h, i32 %still_poison
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2635 store i32 0, i32* %poison_yet_again ; memory at @h[0] is poisoned
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2636
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2637 store i32 %poison, i32* @g ; Poison value stored to memory.
2638 %poison2 = load i32* @g ; Poison value loaded back from memory.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2639
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2640 store volatile i32 %poison, i32* @g ; External observation; undefined behavior.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2641
2642 %narrowaddr = bitcast i32* @g to i16*
2643 %wideaddr = bitcast i32* @g to i64*
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]2644 %poison3 = load i16* %narrowaddr ; Returns a poison value.
2645 %poison4 = load i64* %wideaddr ; Returns a poison value.
Dan Gohman34b3d992010-04-28 00:49:41 +0000[diff] [blame]2646
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2647 %cmp = icmp slt i32 %poison, 0 ; Returns a poison value.
2648 br i1 %cmp, label %true, label %end ; Branch to either destination.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2649
2650true:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2651 store volatile i32 0, i32* @g ; This is control-dependent on %cmp, so
2652 ; it has undefined behavior.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2653 br label %end
2654
2655end:
2656 %p = phi i32 [ 0, %entry ], [ 1, %true ]
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2657 ; Both edges into this PHI are
2658 ; control-dependent on %cmp, so this
2659 ; always results in a poison value.
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2660
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2661 store volatile i32 0, i32* @g ; This would depend on the store in %true
2662 ; if %cmp is true, or the store in %entry
2663 ; otherwise, so this is undefined behavior.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2664
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2665 br i1 %cmp, label %second_true, label %second_end
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2666 ; The same branch again, but this time the
2667 ; true block doesn't have side effects.
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2668
2669second_true:
2670 ; No side effects!
Nick Lewycky64f9fb12011-05-16 19:29:30 +0000[diff] [blame]2671 ret void
Dan Gohmanca4cac42011-04-12 23:05:59 +0000[diff] [blame]2672
2673second_end:
Dan Gohman5cdc51e2011-12-06 03:31:14 +0000[diff] [blame]2674 store volatile i32 0, i32* @g ; This time, the instruction always depends
2675 ; on the store in %end. Also, it is
2676 ; control-equivalent to %end, so this is
Dan Gohmane1a29842011-12-06 03:35:58 +0000[diff] [blame]2677 ; well-defined (ignoring earlier undefined
2678 ; behavior in this example).
Dan Gohmanae11c3f2010-04-26 23:36:52 +0000[diff] [blame]2679</pre>
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2680
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]2681</div>
2682
2683<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2684<h3>
2685 <a name="blockaddress">Addresses of Basic Blocks</a>
2686</h3>
2687
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2688<div>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2689
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2690<p><b><tt>blockaddress(@function, %block)</tt></b></p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2691
2692<p>The '<tt>blockaddress</tt>' constant computes the address of the specified
Chris Lattner2dfdf2a2009-10-27 21:49:40 +0000[diff] [blame]2693 basic block in the specified function, and always has an i8* type. Taking
Chris Lattnercdfc9402009-11-01 01:27:45 +0000[diff] [blame]2694 the address of the entry block is illegal.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2695
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2696<p>This value only has defined behavior when used as an operand to the
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2697 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>' instruction, or for
2698 comparisons against null. Pointer equality tests between labels addresses
2699 results in undefined behavior &mdash; though, again, comparison against null
2700 is ok, and no label is equal to the null pointer. This may be passed around
2701 as an opaque pointer sized value as long as the bits are not inspected. This
2702 allows <tt>ptrtoint</tt> and arithmetic to be performed on these values so
2703 long as the original value is reconstituted before the <tt>indirectbr</tt>
2704 instruction.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]2705
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2706<p>Finally, some targets may provide defined semantics when using the value as
2707 the operand to an inline assembly, but that is target specific.</p>
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]2708
2709</div>
2710
2711
2712<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2713<h3>
2714 <a name="constantexprs">Constant Expressions</a>
2715</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2716
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2717<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2718
2719<p>Constant expressions are used to allow expressions involving other constants
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2720 to be used as constants. Constant expressions may be of
2721 any <a href="#t_firstclass">first class</a> type and may involve any LLVM
2722 operation that does not have side effects (e.g. load and call are not
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]2723 supported). The following is the syntax for constant expressions:</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2724
2725<dl>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2726 <dt><b><tt>trunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2727 <dd>Truncate a constant to another type. The bit size of CST must be larger
2728 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2729
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2730 <dt><b><tt>zext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2731 <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]2732 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2733
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2734 <dt><b><tt>sext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2735 <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]2736 smaller than the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2737
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2738 <dt><b><tt>fptrunc (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2739 <dd>Truncate a floating point constant to another floating point type. The
2740 size of CST must be larger than the size of TYPE. Both types must be
2741 floating point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2742
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2743 <dt><b><tt>fpext (CST to TYPE)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2744 <dd>Floating point extend a constant to another type. The size of CST must be
2745 smaller or equal to the size of TYPE. Both types must be floating
2746 point.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2747
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2748 <dt><b><tt>fptoui (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2749 <dd>Convert a floating point constant to the corresponding unsigned integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2750 constant. TYPE must be a scalar or vector integer type. CST must be of
2751 scalar or vector floating point type. Both CST and TYPE must be scalars,
2752 or vectors of the same number of elements. If the value won't fit in the
2753 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2754
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2755 <dt><b><tt>fptosi (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2756 <dd>Convert a floating point constant to the corresponding signed integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2757 constant. TYPE must be a scalar or vector integer type. CST must be of
2758 scalar or vector floating point type. Both CST and TYPE must be scalars,
2759 or vectors of the same number of elements. If the value won't fit in the
2760 integer type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2761
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2762 <dt><b><tt>uitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2763 <dd>Convert an unsigned integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2764 constant. TYPE must be a scalar or vector floating point type. CST must be
2765 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2766 vectors of the same number of elements. If the value won't fit in the
2767 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2768
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2769 <dt><b><tt>sitofp (CST to TYPE)</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2770 <dd>Convert a signed integer constant to the corresponding floating point
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2771 constant. TYPE must be a scalar or vector floating point type. CST must be
2772 of scalar or vector integer type. Both CST and TYPE must be scalars, or
2773 vectors of the same number of elements. If the value won't fit in the
2774 floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2775
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2776 <dt><b><tt>ptrtoint (CST to TYPE)</tt></b></dt>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2777 <dd>Convert a pointer typed constant to the corresponding integer constant
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2778 <tt>TYPE</tt> must be an integer type. <tt>CST</tt> must be of pointer
2779 type. The <tt>CST</tt> value is zero extended, truncated, or unchanged to
2780 make it fit in <tt>TYPE</tt>.</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2781
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2782 <dt><b><tt>inttoptr (CST to TYPE)</tt></b></dt>
Sylvestre Ledru7f7390e2012-07-25 22:01:31 +0000[diff] [blame]2783 <dd>Convert an integer constant to a pointer constant. TYPE must be a pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2784 type. CST must be of integer type. The CST value is zero extended,
2785 truncated, or unchanged to make it fit in a pointer size. This one is
2786 <i>really</i> dangerous!</dd>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]2787
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2788 <dt><b><tt>bitcast (CST to TYPE)</tt></b></dt>
Chris Lattner03bbad62009-02-28 18:27:03 +0000[diff] [blame]2789 <dd>Convert a constant, CST, to another TYPE. The constraints of the operands
2790 are the same as those for the <a href="#i_bitcast">bitcast
2791 instruction</a>.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2792
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2793 <dt><b><tt>getelementptr (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
2794 <dt><b><tt>getelementptr inbounds (CSTPTR, IDX0, IDX1, ...)</tt></b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2795 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2796 constants. As with the <a href="#i_getelementptr">getelementptr</a>
2797 instruction, the index list may have zero or more indexes, which are
2798 required to make sense for the type of "CSTPTR".</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2799
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2800 <dt><b><tt>select (COND, VAL1, VAL2)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2801 <dd>Perform the <a href="#i_select">select operation</a> on constants.</dd>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2802
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2803 <dt><b><tt>icmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2804 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
2805
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2806 <dt><b><tt>fcmp COND (VAL1, VAL2)</tt></b></dt>
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]2807 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2808
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2809 <dt><b><tt>extractelement (VAL, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2810 <dd>Perform the <a href="#i_extractelement">extractelement operation</a> on
2811 constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]2812
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2813 <dt><b><tt>insertelement (VAL, ELT, IDX)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2814 <dd>Perform the <a href="#i_insertelement">insertelement operation</a> on
2815 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2816
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2817 <dt><b><tt>shufflevector (VEC1, VEC2, IDXMASK)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2818 <dd>Perform the <a href="#i_shufflevector">shufflevector operation</a> on
2819 constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]2820
Nick Lewycky9e130ce2010-05-29 06:44:15 +0000[diff] [blame]2821 <dt><b><tt>extractvalue (VAL, IDX0, IDX1, ...)</tt></b></dt>
2822 <dd>Perform the <a href="#i_extractvalue">extractvalue operation</a> on
2823 constants. The index list is interpreted in a similar manner as indices in
2824 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2825 index value must be specified.</dd>
2826
2827 <dt><b><tt>insertvalue (VAL, ELT, IDX0, IDX1, ...)</tt></b></dt>
2828 <dd>Perform the <a href="#i_insertvalue">insertvalue operation</a> on
2829 constants. The index list is interpreted in a similar manner as indices in
2830 a '<a href="#i_getelementptr">getelementptr</a>' operation. At least one
2831 index value must be specified.</dd>
2832
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]2833 <dt><b><tt>OPCODE (LHS, RHS)</tt></b></dt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2834 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
2835 be any of the <a href="#binaryops">binary</a>
2836 or <a href="#bitwiseops">bitwise binary</a> operations. The constraints
2837 on operands are the same as those for the corresponding instruction
2838 (e.g. no bitwise operations on floating point values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2839</dl>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2840
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]2841</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]2842
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2843</div>
2844
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2845<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2846<h2><a name="othervalues">Other Values</a></h2>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2847<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2848<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2849<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2850<h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2851<a name="inlineasm">Inline Assembler Expressions</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2852</h3>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2853
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2854<div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2855
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2856<p>LLVM supports inline assembler expressions (as opposed
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2857 to <a href="#moduleasm">Module-Level Inline Assembly</a>) through the use of
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2858 a special value. This value represents the inline assembler as a string
2859 (containing the instructions to emit), a list of operand constraints (stored
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2860 as a string), a flag that indicates whether or not the inline asm
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2861 expression has side effects, and a flag indicating whether the function
2862 containing the asm needs to align its stack conservatively. An example
2863 inline assembler expression is:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2864
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2865<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2866i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2867</pre>
2868
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2869<p>Inline assembler expressions may <b>only</b> be used as the callee operand of
Nuno Lopes0b5f1ac2012-07-17 23:51:33 +0000[diff] [blame]2870 a <a href="#i_call"><tt>call</tt></a> or an
2871 <a href="#i_invoke"><tt>invoke</tt></a> instruction.
2872 Thus, typically we have:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2873
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2874<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2875%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]2876</pre>
2877
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2878<p>Inline asms with side effects not visible in the constraint list must be
2879 marked as having side effects. This is done through the use of the
2880 '<tt>sideeffect</tt>' keyword, like so:</p>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2881
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2882<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2883call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2884</pre>
2885
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2886<p>In some cases inline asms will contain code that will not work unless the
2887 stack is aligned in some way, such as calls or SSE instructions on x86,
2888 yet will not contain code that does that alignment within the asm.
2889 The compiler should make conservative assumptions about what the asm might
2890 contain and should generate its usual stack alignment code in the prologue
2891 if the '<tt>alignstack</tt>' keyword is present:</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2892
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2893<pre class="doc_code">
Dale Johannesen8ba2d5b2009-10-21 23:28:00 +0000[diff] [blame]2894call void asm alignstack "eieio", ""()
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2895</pre>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2896
Chad Rosier36547342012-09-05 00:08:17 +0000[diff] [blame]2897<p>Inline asms also support using non-standard assembly dialects. The standard
2898 dialect is ATT, which is assumed when the '<tt>nsdialect</tt>' keyword is not
2899 present. When the '<tt>nsdialect</tt>' keyword is present, the dialect is
2900 assumed to be Intel. Currently, ATT and Intel are the only supported
2901 dialects. An example is:</p>
2902
2903<pre class="doc_code">
2904call void asm nsdialect "eieio", ""()
2905</pre>
2906
2907<p>If multiple keywords appear the '<tt>sideeffect</tt>' keyword must come
2908 first, the '<tt>alignstack</tt>' keyword second and the
2909 '<tt>nsdialect</tt>' keyword last.</p>
Dale Johannesen09fed252009-10-13 21:56:55 +0000[diff] [blame]2910
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2911<!--
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]2912<p>TODO: The format of the asm and constraints string still need to be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]2913 documented here. Constraints on what can be done (e.g. duplication, moving,
2914 etc need to be documented). This is probably best done by reference to
2915 another document that covers inline asm from a holistic perspective.</p>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2916 -->
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2917
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2918<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2919<h4>
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2920 <a name="inlineasm_md">Inline Asm Metadata</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2921</h4>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2922
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2923<div>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2924
Bill Wendlingaee0f452011-11-30 21:52:43 +0000[diff] [blame]2925<p>The call instructions that wrap inline asm nodes may have a
2926 "<tt>!srcloc</tt>" MDNode attached to it that contains a list of constant
2927 integers. If present, the code generator will use the integer as the
2928 location cookie value when report errors through the <tt>LLVMContext</tt>
2929 error reporting mechanisms. This allows a front-end to correlate backend
2930 errors that occur with inline asm back to the source code that produced it.
2931 For example:</p>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2932
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]2933<pre class="doc_code">
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2934call void asm sideeffect "something bad", ""()<b>, !srcloc !42</b>
2935...
2936!42 = !{ i32 1234567 }
2937</pre>
Chris Lattnercf9a4152010-04-07 05:38:05 +0000[diff] [blame]2938
2939<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]2940 IR. If the MDNode contains multiple constants, the code generator will use
Chris Lattnerce1b9ad2010-11-17 08:20:42 +0000[diff] [blame]2941 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]2942
2943</div>
2944
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2945</div>
2946
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2947<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]2948<h3>
2949 <a name="metadata">Metadata Nodes and Metadata Strings</a>
2950</h3>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2951
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]2952<div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2953
2954<p>LLVM IR allows metadata to be attached to instructions in the program that
2955 can convey extra information about the code to the optimizers and code
2956 generator. One example application of metadata is source-level debug
2957 information. There are two metadata primitives: strings and nodes. All
2958 metadata has the <tt>metadata</tt> type and is identified in syntax by a
2959 preceding exclamation point ('<tt>!</tt>').</p>
2960
2961<p>A metadata string is a string surrounded by double quotes. It can contain
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2962 any character by escaping non-printable characters with "<tt>\xx</tt>" where
2963 "<tt>xx</tt>" is the two digit hex code. For example:
2964 "<tt>!"test\00"</tt>".</p>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2965
2966<p>Metadata nodes are represented with notation similar to structure constants
2967 (a comma separated list of elements, surrounded by braces and preceded by an
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2968 exclamation point). Metadata nodes can have any values as their operand. For
2969 example:</p>
2970
2971<div class="doc_code">
2972<pre>
2973!{ metadata !"test\00", i32 10}
2974</pre>
2975</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2976
2977<p>A <a href="#namedmetadatastructure">named metadata</a> is a collection of
2978 metadata nodes, which can be looked up in the module symbol table. For
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2979 example:</p>
2980
2981<div class="doc_code">
2982<pre>
2983!foo = metadata !{!4, !3}
2984</pre>
2985</div>
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]2986
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2987<p>Metadata can be used as function arguments. Here <tt>llvm.dbg.value</tt>
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2988 function is using two metadata arguments:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2989
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]2990<div class="doc_code">
2991<pre>
2992call void @llvm.dbg.value(metadata !24, i64 0, metadata !25)
2993</pre>
2994</div>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2995
2996<p>Metadata can be attached with an instruction. Here metadata <tt>!21</tt> is
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]2997 attached to the <tt>add</tt> instruction using the <tt>!dbg</tt>
2998 identifier:</p>
Devang Patele1d50cd2010-03-04 23:44:48 +0000[diff] [blame]2999
Bill Wendling9ff5de92011-03-02 02:17:11 +0000[diff] [blame]3000<div class="doc_code">
3001<pre>
3002%indvar.next = add i64 %indvar, 1, !dbg !21
3003</pre>
3004</div>
3005
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]3006<p>More information about specific metadata nodes recognized by the optimizers
3007 and code generator is found below.</p>
3008
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]3009<!-- _______________________________________________________________________ -->
Peter Collingbourne249d9532011-10-27 19:19:07 +0000[diff] [blame]3010<h4>
3011 <a name="tbaa">'<tt>tbaa</tt>' Metadata</a>
3012</h4>
3013
3014<div>
3015
3016<p>In LLVM IR, memory does not have types, so LLVM's own type system is not
3017 suitable for doing TBAA. Instead, metadata is added to the IR to describe
3018 a type system of a higher level language. This can be used to implement
3019 typical C/C++ TBAA, but it can also be used to implement custom alias
3020 analysis behavior for other languages.</p>
3021
3022<p>The current metadata format is very simple. TBAA metadata nodes have up to
3023 three fields, e.g.:</p>
3024
3025<div class="doc_code">
3026<pre>
3027!0 = metadata !{ metadata !"an example type tree" }
3028!1 = metadata !{ metadata !"int", metadata !0 }
3029!2 = metadata !{ metadata !"float", metadata !0 }
3030!3 = metadata !{ metadata !"const float", metadata !2, i64 1 }
3031</pre>
3032</div>
3033
3034<p>The first field is an identity field. It can be any value, usually
3035 a metadata string, which uniquely identifies the type. The most important
3036 name in the tree is the name of the root node. Two trees with
3037 different root node names are entirely disjoint, even if they
3038 have leaves with common names.</p>
3039
3040<p>The second field identifies the type's parent node in the tree, or
3041 is null or omitted for a root node. A type is considered to alias
3042 all of its descendants and all of its ancestors in the tree. Also,
3043 a type is considered to alias all types in other trees, so that
3044 bitcode produced from multiple front-ends is handled conservatively.</p>
3045
3046<p>If the third field is present, it's an integer which if equal to 1
3047 indicates that the type is "constant" (meaning
3048 <tt>pointsToConstantMemory</tt> should return true; see
3049 <a href="AliasAnalysis.html#OtherItfs">other useful
3050 <tt>AliasAnalysis</tt> methods</a>).</p>
3051
3052</div>
3053
Bill Wendlingf6cc4c22011-11-30 21:43:43 +0000[diff] [blame]3054<!-- _______________________________________________________________________ -->
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3055<h4>
Duncan Sands5e5c5f82012-04-14 12:36:06 +0000[diff] [blame]3056 <a name="fpmath">'<tt>fpmath</tt>' Metadata</a>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3057</h4>
3058
3059<div>
3060
Duncan Sands5e5c5f82012-04-14 12:36:06 +0000[diff] [blame]3061<p><tt>fpmath</tt> metadata may be attached to any instruction of floating point
Duncan Sands8883c432012-04-16 16:28:59 +0000[diff] [blame]3062 type. It can be used to express the maximum acceptable error in the result of
3063 that instruction, in ULPs, thus potentially allowing the compiler to use a
Duncan Sands2867c852012-04-16 19:39:33 +0000[diff] [blame]3064 more efficient but less accurate method of computing it. ULP is defined as
3065 follows:</p>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3066
Bill Wendling0656e252011-11-09 19:33:56 +0000[diff] [blame]3067<blockquote>
3068
3069<p>If <tt>x</tt> is a real number that lies between two finite consecutive
3070 floating-point numbers <tt>a</tt> and <tt>b</tt>, without being equal to one
3071 of them, then <tt>ulp(x) = |b - a|</tt>, otherwise <tt>ulp(x)</tt> is the
3072 distance between the two non-equal finite floating-point numbers nearest
3073 <tt>x</tt>. Moreover, <tt>ulp(NaN)</tt> is <tt>NaN</tt>.</p>
3074
3075</blockquote>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3076
Duncan Sands8883c432012-04-16 16:28:59 +0000[diff] [blame]3077<p>The metadata node shall consist of a single positive floating point number
Duncan Sands2867c852012-04-16 19:39:33 +0000[diff] [blame]3078 representing the maximum relative error, for example:</p>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3079
3080<div class="doc_code">
3081<pre>
Duncan Sands8883c432012-04-16 16:28:59 +0000[diff] [blame]3082!0 = metadata !{ float 2.5 } ; maximum acceptable inaccuracy is 2.5 ULPs
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3083</pre>
3084</div>
3085
NAKAMURA Takumi9c55f592012-03-27 11:25:16 +0000[diff] [blame]3086</div>
3087
Rafael Espindola39dd3282012-03-24 00:14:51 +0000[diff] [blame]3088<!-- _______________________________________________________________________ -->
3089<h4>
3090 <a name="range">'<tt>range</tt>' Metadata</a>
3091</h4>
3092
3093<div>
3094<p><tt>range</tt> metadata may be attached only to loads of integer types. It
3095 expresses the possible ranges the loaded value is in. The ranges are
3096 represented with a flattened list of integers. The loaded value is known to
3097 be in the union of the ranges defined by each consecutive pair. Each pair
3098 has the following properties:</p>
3099<ul>
3100 <li>The type must match the type loaded by the instruction.</li>
3101 <li>The pair <tt>a,b</tt> represents the range <tt>[a,b)</tt>.</li>
3102 <li>Both <tt>a</tt> and <tt>b</tt> are constants.</li>
3103 <li>The range is allowed to wrap.</li>
3104 <li>The range should not represent the full or empty set. That is,
3105 <tt>a!=b</tt>. </li>
3106</ul>
Rafael Espindolaa1b95f52012-05-31 16:04:26 +0000[diff] [blame]3107<p> In addition, the pairs must be in signed order of the lower bound and
3108 they must be non-contiguous.</p>
Rafael Espindola39dd3282012-03-24 00:14:51 +0000[diff] [blame]3109
3110<p>Examples:</p>
3111<div class="doc_code">
3112<pre>
3113 %a = load i8* %x, align 1, !range !0 ; Can only be 0 or 1
3114 %b = load i8* %y, align 1, !range !1 ; Can only be 255 (-1), 0 or 1
3115 %c = load i8* %z, align 1, !range !2 ; Can only be 0, 1, 3, 4 or 5
Rafael Espindolac49b29e2012-05-31 13:45:46 +0000[diff] [blame]3116 %d = load i8* %z, align 1, !range !3 ; Can only be -2, -1, 3, 4 or 5
Rafael Espindola39dd3282012-03-24 00:14:51 +0000[diff] [blame]3117...
3118!0 = metadata !{ i8 0, i8 2 }
3119!1 = metadata !{ i8 255, i8 2 }
3120!2 = metadata !{ i8 0, i8 2, i8 3, i8 6 }
Rafael Espindolac49b29e2012-05-31 13:45:46 +0000[diff] [blame]3121!3 = metadata !{ i8 -2, i8 0, i8 3, i8 6 }
Rafael Espindola39dd3282012-03-24 00:14:51 +0000[diff] [blame]3122</pre>
3123</div>
3124</div>
Peter Collingbourne999f90b2011-10-27 19:19:14 +0000[diff] [blame]3125</div>
3126
Chris Lattnere6a5ddd2010-01-15 21:50:19 +0000[diff] [blame]3127</div>
3128
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3129<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3130<h2>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3131 <a name="module_flags">Module Flags Metadata</a>
3132</h2>
3133<!-- *********************************************************************** -->
3134
3135<div>
3136
3137<p>Information about the module as a whole is difficult to convey to LLVM's
3138 subsystems. The LLVM IR isn't sufficient to transmit this
3139 information. The <tt>llvm.module.flags</tt> named metadata exists in order to
3140 facilitate this. These flags are in the form of key / value pairs &mdash;
3141 much like a dictionary &mdash; making it easy for any subsystem who cares
3142 about a flag to look it up.</p>
3143
3144<p>The <tt>llvm.module.flags</tt> metadata contains a list of metadata
3145 triplets. Each triplet has the following form:</p>
3146
3147<ul>
3148 <li>The first element is a <i>behavior</i> flag, which specifies the behavior
3149 when two (or more) modules are merged together, and it encounters two (or
3150 more) metadata with the same ID. The supported behaviors are described
3151 below.</li>
3152
3153 <li>The second element is a metadata string that is a unique ID for the
3154 metadata. How each ID is interpreted is documented below.</li>
3155
3156 <li>The third element is the value of the flag.</li>
3157</ul>
3158
3159<p>When two (or more) modules are merged together, the resulting
3160 <tt>llvm.module.flags</tt> metadata is the union of the
3161 modules' <tt>llvm.module.flags</tt> metadata. The only exception being a flag
3162 with the <i>Override</i> behavior, which may override another flag's value
3163 (see below).</p>
3164
3165<p>The following behaviors are supported:</p>
3166
3167<table border="1" cellspacing="0" cellpadding="4">
3168 <tbody>
3169 <tr>
3170 <th>Value</th>
3171 <th>Behavior</th>
3172 </tr>
3173 <tr>
3174 <td>1</td>
3175 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3176 <dl>
3177 <dt><b>Error</b></dt>
3178 <dd>Emits an error if two values disagree. It is an error to have an ID
3179 with both an Error and a Warning behavior.</dd>
3180 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3181 </td>
3182 </tr>
3183 <tr>
3184 <td>2</td>
3185 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3186 <dl>
3187 <dt><b>Warning</b></dt>
3188 <dd>Emits a warning if two values disagree.</dd>
3189 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3190 </td>
3191 </tr>
3192 <tr>
3193 <td>3</td>
3194 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3195 <dl>
3196 <dt><b>Require</b></dt>
3197 <dd>Emits an error when the specified value is not present or doesn't
3198 have the specified value. It is an error for two (or more)
3199 <tt>llvm.module.flags</tt> with the same ID to have the Require
3200 behavior but different values. There may be multiple Require flags
3201 per ID.</dd>
3202 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3203 </td>
3204 </tr>
3205 <tr>
3206 <td>4</td>
3207 <td align="left">
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3208 <dl>
3209 <dt><b>Override</b></dt>
3210 <dd>Uses the specified value if the two values disagree. It is an
3211 error for two (or more) <tt>llvm.module.flags</tt> with the same
3212 ID to have the Override behavior but different values.</dd>
3213 </dl>
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3214 </td>
3215 </tr>
3216 </tbody>
3217</table>
3218
3219<p>An example of module flags:</p>
3220
3221<pre class="doc_code">
3222!0 = metadata !{ i32 1, metadata !"foo", i32 1 }
3223!1 = metadata !{ i32 4, metadata !"bar", i32 37 }
3224!2 = metadata !{ i32 2, metadata !"qux", i32 42 }
3225!3 = metadata !{ i32 3, metadata !"qux",
3226 metadata !{
3227 metadata !"foo", i32 1
3228 }
3229}
3230!llvm.module.flags = !{ !0, !1, !2, !3 }
3231</pre>
3232
3233<ul>
3234 <li><p>Metadata <tt>!0</tt> has the ID <tt>!"foo"</tt> and the value '1'. The
3235 behavior if two or more <tt>!"foo"</tt> flags are seen is to emit an
3236 error if their values are not equal.</p></li>
3237
3238 <li><p>Metadata <tt>!1</tt> has the ID <tt>!"bar"</tt> and the value '37'. The
3239 behavior if two or more <tt>!"bar"</tt> flags are seen is to use the
3240 value '37' if their values are not equal.</p></li>
3241
3242 <li><p>Metadata <tt>!2</tt> has the ID <tt>!"qux"</tt> and the value '42'. The
3243 behavior if two or more <tt>!"qux"</tt> flags are seen is to emit a
3244 warning if their values are not equal.</p></li>
3245
3246 <li><p>Metadata <tt>!3</tt> has the ID <tt>!"qux"</tt> and the value:</p>
3247
3248<pre class="doc_code">
3249metadata !{ metadata !"foo", i32 1 }
3250</pre>
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3251
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3252 <p>The behavior is to emit an error if the <tt>llvm.module.flags</tt> does
3253 not contain a flag with the ID <tt>!"foo"</tt> that has the value
3254 '1'. If two or more <tt>!"qux"</tt> flags exist, then they must have
3255 the same value or an error will be issued.</p></li>
3256</ul>
3257
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3258
3259<!-- ======================================================================= -->
3260<h3>
3261<a name="objc_gc_flags">Objective-C Garbage Collection Module Flags Metadata</a>
3262</h3>
3263
3264<div>
3265
3266<p>On the Mach-O platform, Objective-C stores metadata about garbage collection
3267 in a special section called "image info". The metadata consists of a version
3268 number and a bitmask specifying what types of garbage collection are
3269 supported (if any) by the file. If two or more modules are linked together
3270 their garbage collection metadata needs to be merged rather than appended
3271 together.</p>
3272
3273<p>The Objective-C garbage collection module flags metadata consists of the
3274 following key-value pairs:</p>
3275
3276<table border="1" cellspacing="0" cellpadding="4">
Bill Wendlingb3ef2232012-03-06 09:23:25 +0000[diff] [blame]3277 <col width="30%">
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3278 <tbody>
3279 <tr>
Bill Wendlinga0edecf2012-03-06 09:17:04 +0000[diff] [blame]3280 <th>Key</th>
Bill Wendlingf7b367c2012-02-16 01:10:50 +0000[diff] [blame]3281 <th>Value</th>
3282 </tr>
3283 <tr>
3284 <td><tt>Objective-C&nbsp;Version</tt></td>
3285 <td align="left"><b>[Required]</b> &mdash; The Objective-C ABI
3286 version. Valid values are 1 and 2.</td>
3287 </tr>
3288 <tr>
3289 <td><tt>Objective-C&nbsp;Image&nbsp;Info&nbsp;Version</tt></td>
3290 <td align="left"><b>[Required]</b> &mdash; The version of the image info
3291 section. Currently always 0.</td>
3292 </tr>
3293 <tr>
3294 <td><tt>Objective-C&nbsp;Image&nbsp;Info&nbsp;Section</tt></td>
3295 <td align="left"><b>[Required]</b> &mdash; The section to place the
3296 metadata. Valid values are <tt>"__OBJC, __image_info, regular"</tt> for
3297 Objective-C ABI version 1, and <tt>"__DATA,__objc_imageinfo, regular,
3298 no_dead_strip"</tt> for Objective-C ABI version 2.</td>
3299 </tr>
3300 <tr>
3301 <td><tt>Objective-C&nbsp;Garbage&nbsp;Collection</tt></td>
3302 <td align="left"><b>[Required]</b> &mdash; Specifies whether garbage
3303 collection is supported or not. Valid values are 0, for no garbage
3304 collection, and 2, for garbage collection supported.</td>
3305 </tr>
3306 <tr>
3307 <td><tt>Objective-C&nbsp;GC&nbsp;Only</tt></td>
3308 <td align="left"><b>[Optional]</b> &mdash; Specifies that only garbage
3309 collection is supported. If present, its value must be 6. This flag
3310 requires that the <tt>Objective-C Garbage Collection</tt> flag have the
3311 value 2.</td>
3312 </tr>
3313 </tbody>
3314</table>
3315
3316<p>Some important flag interactions:</p>
3317
3318<ul>
3319 <li>If a module with <tt>Objective-C Garbage Collection</tt> set to 0 is
3320 merged with a module with <tt>Objective-C Garbage Collection</tt> set to
3321 2, then the resulting module has the <tt>Objective-C Garbage
3322 Collection</tt> flag set to 0.</li>
3323
3324 <li>A module with <tt>Objective-C Garbage Collection</tt> set to 0 cannot be
3325 merged with a module with <tt>Objective-C GC Only</tt> set to 6.</li>
3326</ul>
3327
3328</div>
3329
Bill Wendlingb9d75a92012-02-11 11:59:36 +0000[diff] [blame]3330</div>
3331
3332<!-- *********************************************************************** -->
3333<h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3334 <a name="intrinsic_globals">Intrinsic Global Variables</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3335</h2>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3336<!-- *********************************************************************** -->
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3337<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3338<p>LLVM has a number of "magic" global variables that contain data that affect
3339code generation or other IR semantics. These are documented here. All globals
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3340of this sort should have a section specified as "<tt>llvm.metadata</tt>". This
3341section and all globals that start with "<tt>llvm.</tt>" are reserved for use
3342by LLVM.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3343
3344<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3345<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3346<a name="intg_used">The '<tt>llvm.used</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3347</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3348
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3349<div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3350
3351<p>The <tt>@llvm.used</tt> global is an array with i8* element type which has <a
3352href="#linkage_appending">appending linkage</a>. This array contains a list of
3353pointers to global variables and functions which may optionally have a pointer
3354cast formed of bitcast or getelementptr. For example, a legal use of it is:</p>
3355
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3356<div class="doc_code">
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3357<pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3358@X = global i8 4
3359@Y = global i32 123
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3360
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3361@llvm.used = appending global [2 x i8*] [
3362 i8* @X,
3363 i8* bitcast (i32* @Y to i8*)
3364], section "llvm.metadata"
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3365</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3366</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3367
3368<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]3369 compiler, assembler, and linker are required to treat the symbol as if there
3370 is a reference to the global that it cannot see. For example, if a variable
3371 has internal linkage and no references other than that from
3372 the <tt>@llvm.used</tt> list, it cannot be deleted. This is commonly used to
3373 represent references from inline asms and other things the compiler cannot
3374 "see", and corresponds to "<tt>attribute((used))</tt>" in GNU C.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3375
3376<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]3377 object file to prevent the assembler and linker from molesting the
3378 symbol.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3379
3380</div>
3381
3382<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3383<h3>
3384 <a name="intg_compiler_used">
3385 The '<tt>llvm.compiler.used</tt>' Global Variable
3386 </a>
3387</h3>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3388
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3389<div>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3390
3391<p>The <tt>@llvm.compiler.used</tt> directive is the same as the
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3392 <tt>@llvm.used</tt> directive, except that it only prevents the compiler from
3393 touching the symbol. On targets that support it, this allows an intelligent
3394 linker to optimize references to the symbol without being impeded as it would
3395 be by <tt>@llvm.used</tt>.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3396
3397<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]3398 should not be exposed to source languages.</p>
Chris Lattner401e10c2009-07-20 06:14:25 +0000[diff] [blame]3399
3400</div>
3401
3402<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3403<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3404<a name="intg_global_ctors">The '<tt>llvm.global_ctors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3405</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3406
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3407<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3408
3409<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3410<pre>
3411%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3412@llvm.global_ctors = appending global [1 x %0] [%0 { i32 65535, void ()* @ctor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3413</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3414</div>
3415
3416<p>The <tt>@llvm.global_ctors</tt> array contains a list of constructor
3417 functions and associated priorities. The functions referenced by this array
3418 will be called in ascending order of priority (i.e. lowest first) when the
3419 module is loaded. The order of functions with the same priority is not
3420 defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3421
3422</div>
3423
3424<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3425<h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3426<a name="intg_global_dtors">The '<tt>llvm.global_dtors</tt>' Global Variable</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3427</h3>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3428
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3429<div>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3430
3431<div class="doc_code">
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3432<pre>
3433%0 = type { i32, void ()* }
David Chisnall27195a52010-04-30 19:27:35 +0000[diff] [blame]3434@llvm.global_dtors = appending global [1 x %0] [%0 { i32 65535, void ()* @dtor }]
David Chisnalle31e9962010-04-30 19:23:49 +0000[diff] [blame]3435</pre>
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3436</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3437
Bill Wendling9ae75632011-11-08 00:32:45 +0000[diff] [blame]3438<p>The <tt>@llvm.global_dtors</tt> array contains a list of destructor functions
3439 and associated priorities. The functions referenced by this array will be
3440 called in descending order of priority (i.e. highest first) when the module
3441 is loaded. The order of functions with the same priority is not defined.</p>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3442
3443</div>
3444
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3445</div>
Chris Lattner857755c2009-07-20 05:55:19 +0000[diff] [blame]3446
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]3447<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3448<h2><a name="instref">Instruction Reference</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3449<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3450
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3451<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3452
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3453<p>The LLVM instruction set consists of several different classifications of
3454 instructions: <a href="#terminators">terminator
3455 instructions</a>, <a href="#binaryops">binary instructions</a>,
3456 <a href="#bitwiseops">bitwise binary instructions</a>,
3457 <a href="#memoryops">memory instructions</a>, and
3458 <a href="#otherops">other instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3459
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3460<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3461<h3>
3462 <a name="terminators">Terminator Instructions</a>
3463</h3>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3464
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3465<div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3466
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3467<p>As mentioned <a href="#functionstructure">previously</a>, every basic block
3468 in a program ends with a "Terminator" instruction, which indicates which
3469 block should be executed after the current block is finished. These
3470 terminator instructions typically yield a '<tt>void</tt>' value: they produce
3471 control flow, not values (the one exception being the
3472 '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
3473
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3474<p>The terminator instructions are:
3475 '<a href="#i_ret"><tt>ret</tt></a>',
3476 '<a href="#i_br"><tt>br</tt></a>',
3477 '<a href="#i_switch"><tt>switch</tt></a>',
3478 '<a href="#i_indirectbr"><tt>indirectbr</tt></a>',
3479 '<a href="#i_invoke"><tt>invoke</tt></a>',
Chris Lattner6445ecb2011-08-02 20:29:13 +0000[diff] [blame]3480 '<a href="#i_resume"><tt>resume</tt></a>', and
3481 '<a href="#i_unreachable"><tt>unreachable</tt></a>'.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]3482
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3483<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3484<h4>
3485 <a name="i_ret">'<tt>ret</tt>' Instruction</a>
3486</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3487
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3488<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3489
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3490<h5>Syntax:</h5>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3491<pre>
3492 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]3493 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3494</pre>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3495
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3496<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3497<p>The '<tt>ret</tt>' instruction is used to return control flow (and optionally
3498 a value) from a function back to the caller.</p>
3499
3500<p>There are two forms of the '<tt>ret</tt>' instruction: one that returns a
3501 value and then causes control flow, and one that just causes control flow to
3502 occur.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3503
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3504<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3505<p>The '<tt>ret</tt>' instruction optionally accepts a single argument, the
3506 return value. The type of the return value must be a
3507 '<a href="#t_firstclass">first class</a>' type.</p>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]3508
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3509<p>A function is not <a href="#wellformed">well formed</a> if it it has a
3510 non-void return type and contains a '<tt>ret</tt>' instruction with no return
3511 value or a return value with a type that does not match its type, or if it
3512 has a void return type and contains a '<tt>ret</tt>' instruction with a
3513 return value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3514
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3515<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3516<p>When the '<tt>ret</tt>' instruction is executed, control flow returns back to
3517 the calling function's context. If the caller is a
3518 "<a href="#i_call"><tt>call</tt></a>" instruction, execution continues at the
3519 instruction after the call. If the caller was an
3520 "<a href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues at
3521 the beginning of the "normal" destination block. If the instruction returns
3522 a value, that value shall set the call or invoke instruction's return
3523 value.</p>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3524
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3525<h5>Example:</h5>
Chris Lattnerf4cde4e2008-04-23 04:59:35 +0000[diff] [blame]3526<pre>
3527 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]3528 ret void <i>; Return from a void function</i>
Bill Wendling0a4bbbf2009-02-28 22:12:54 +0000[diff] [blame]3529 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]3530</pre>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]3531
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3532</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3533<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3534<h4>
3535 <a name="i_br">'<tt>br</tt>' Instruction</a>
3536</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3537
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3538<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3539
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3540<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3541<pre>
Bill Wendlingb3aa4712011-07-26 10:41:15 +0000[diff] [blame]3542 br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;
3543 br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3544</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3545
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3546<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3547<p>The '<tt>br</tt>' instruction is used to cause control flow to transfer to a
3548 different basic block in the current function. There are two forms of this
3549 instruction, corresponding to a conditional branch and an unconditional
3550 branch.</p>
3551
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3552<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3553<p>The conditional branch form of the '<tt>br</tt>' instruction takes a single
3554 '<tt>i1</tt>' value and two '<tt>label</tt>' values. The unconditional form
3555 of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>' value as a
3556 target.</p>
3557
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3558<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3559<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3560 argument is evaluated. If the value is <tt>true</tt>, control flows to the
3561 '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
3562 control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
3563
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3564<h5>Example:</h5>
Bill Wendlingc39e3e02009-07-20 02:39:26 +0000[diff] [blame]3565<pre>
3566Test:
3567 %cond = <a href="#i_icmp">icmp</a> eq i32 %a, %b
3568 br i1 %cond, label %IfEqual, label %IfUnequal
3569IfEqual:
3570 <a href="#i_ret">ret</a> i32 1
3571IfUnequal:
3572 <a href="#i_ret">ret</a> i32 0
3573</pre>
3574
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3575</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3576
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3577<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3578<h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3579 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3580</h4>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3581
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3582<div>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3583
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3584<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3585<pre>
3586 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
3587</pre>
3588
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3589<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3590<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]3591 several different places. It is a generalization of the '<tt>br</tt>'
3592 instruction, allowing a branch to occur to one of many possible
3593 destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3594
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3595<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3596<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3597 comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination,
3598 and an array of pairs of comparison value constants and '<tt>label</tt>'s.
3599 The table is not allowed to contain duplicate constant entries.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3600
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3601<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3602<p>The <tt>switch</tt> instruction specifies a table of values and
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3603 destinations. When the '<tt>switch</tt>' instruction is executed, this table
3604 is searched for the given value. If the value is found, control flow is
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]3605 transferred to the corresponding destination; otherwise, control flow is
3606 transferred to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3607
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3608<h5>Implementation:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3609<p>Depending on properties of the target machine and the particular
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3610 <tt>switch</tt> instruction, this instruction may be code generated in
3611 different ways. For example, it could be generated as a series of chained
3612 conditional branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3613
3614<h5>Example:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3615<pre>
3616 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3617 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3618 switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3619
3620 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3621 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]3622
3623 <i>; Implement a jump table:</i>
Dan Gohman2a08c532009-01-04 23:44:43 +0000[diff] [blame]3624 switch i32 %val, label %otherwise [ i32 0, label %onzero
3625 i32 1, label %onone
3626 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3627</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3628
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3629</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3630
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3631
3632<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3633<h4>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3634 <a name="i_indirectbr">'<tt>indirectbr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3635</h4>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3636
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3637<div>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3638
3639<h5>Syntax:</h5>
3640<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3641 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]3642</pre>
3643
3644<h5>Overview:</h5>
3645
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3646<p>The '<tt>indirectbr</tt>' instruction implements an indirect branch to a label
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3647 within the current function, whose address is specified by
Chris Lattnerc6f44362009-10-27 21:01:34 +0000[diff] [blame]3648 "<tt>address</tt>". Address must be derived from a <a
3649 href="#blockaddress">blockaddress</a> constant.</p>
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3650
3651<h5>Arguments:</h5>
3652
3653<p>The '<tt>address</tt>' argument is the address of the label to jump to. The
3654 rest of the arguments indicate the full set of possible destinations that the
3655 address may point to. Blocks are allowed to occur multiple times in the
3656 destination list, though this isn't particularly useful.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]3657
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3658<p>This destination list is required so that dataflow analysis has an accurate
3659 understanding of the CFG.</p>
3660
3661<h5>Semantics:</h5>
3662
3663<p>Control transfers to the block specified in the address argument. All
3664 possible destination blocks must be listed in the label list, otherwise this
3665 instruction has undefined behavior. This implies that jumps to labels
3666 defined in other functions have undefined behavior as well.</p>
3667
3668<h5>Implementation:</h5>
3669
3670<p>This is typically implemented with a jump through a register.</p>
3671
3672<h5>Example:</h5>
3673<pre>
Chris Lattnerab21db72009-10-28 00:19:10 +0000[diff] [blame]3674 indirectbr i8* %Addr, [ label %bb1, label %bb2, label %bb3 ]
Chris Lattnerf9be95f2009-10-27 19:13:16 +0000[diff] [blame]3675</pre>
3676
3677</div>
3678
3679
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3680<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3681<h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3682 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3683</h4>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3684
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3685<div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3686
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3687<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3688<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3689 &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]3690 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3691</pre>
3692
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3693<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3694<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3695 function, with the possibility of control flow transfer to either the
3696 '<tt>normal</tt>' label or the '<tt>exception</tt>' label. If the callee
3697 function returns with the "<tt><a href="#i_ret">ret</a></tt>" instruction,
3698 control flow will return to the "normal" label. If the callee (or any
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]3699 indirect callees) returns via the "<a href="#i_resume"><tt>resume</tt></a>"
3700 instruction or other exception handling mechanism, control is interrupted and
3701 continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3702
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3703<p>The '<tt>exception</tt>' label is a
3704 <i><a href="ExceptionHandling.html#overview">landing pad</a></i> for the
3705 exception. As such, '<tt>exception</tt>' label is required to have the
3706 "<a href="#i_landingpad"><tt>landingpad</tt></a>" instruction, which contains
Chad Rosier85f5a1a2011-12-09 02:00:44 +0000[diff] [blame]3707 the information about the behavior of the program after unwinding
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3708 happens, as its first non-PHI instruction. The restrictions on the
3709 "<tt>landingpad</tt>" instruction's tightly couples it to the
3710 "<tt>invoke</tt>" instruction, so that the important information contained
3711 within the "<tt>landingpad</tt>" instruction can't be lost through normal
3712 code motion.</p>
3713
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3714<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3715<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3716
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3717<ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3718 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
3719 convention</a> the call should use. If none is specified, the call
3720 defaults to using C calling conventions.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3721
3722 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3723 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
3724 '<tt>inreg</tt>' attributes are valid here.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]3725
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3726 <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]3727 function value being invoked. In most cases, this is a direct function
3728 invocation, but indirect <tt>invoke</tt>s are just as possible, branching
3729 off an arbitrary pointer to function value.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3730
3731 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3732 function to be invoked. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3733
3734 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]3735 signature argument types and parameter attributes. All arguments must be
3736 of <a href="#t_firstclass">first class</a> type. If the function
3737 signature indicates the function accepts a variable number of arguments,
3738 the extra arguments can be specified.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3739
3740 <li>'<tt>normal label</tt>': the label reached when the called function
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3741 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3742
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]3743 <li>'<tt>exception label</tt>': the label reached when a callee returns via
3744 the <a href="#i_resume"><tt>resume</tt></a> instruction or other exception
3745 handling mechanism.</li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3746
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]3747 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3748 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
3749 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3750</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3751
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3752<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3753<p>This instruction is designed to operate as a standard
3754 '<tt><a href="#i_call">call</a></tt>' instruction in most regards. The
3755 primary difference is that it establishes an association with a label, which
3756 is used by the runtime library to unwind the stack.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3757
3758<p>This instruction is used in languages with destructors to ensure that proper
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3759 cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
3760 exception. Additionally, this is important for implementation of
3761 '<tt>catch</tt>' clauses in high-level languages that support them.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3762
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3763<p>For the purposes of the SSA form, the definition of the value returned by the
3764 '<tt>invoke</tt>' instruction is deemed to occur on the edge from the current
3765 block to the "normal" label. If the callee unwinds then no return value is
3766 available.</p>
Dan Gohmanf96a4992009-05-22 21:47:08 +0000[diff] [blame]3767
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3768<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3769<pre>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3770 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3771 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewyckyd703f652008-03-16 07:18:12 +0000[diff] [blame]3772 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]3773 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3774</pre>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3775
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3776</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3777
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3778 <!-- _______________________________________________________________________ -->
3779
3780<h4>
3781 <a name="i_resume">'<tt>resume</tt>' Instruction</a>
3782</h4>
3783
3784<div>
3785
3786<h5>Syntax:</h5>
3787<pre>
3788 resume &lt;type&gt; &lt;value&gt;
3789</pre>
3790
3791<h5>Overview:</h5>
3792<p>The '<tt>resume</tt>' instruction is a terminator instruction that has no
3793 successors.</p>
3794
3795<h5>Arguments:</h5>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3796<p>The '<tt>resume</tt>' instruction requires one argument, which must have the
Bill Wendlinge4ad50b2011-08-03 18:37:32 +0000[diff] [blame]3797 same type as the result of any '<tt>landingpad</tt>' instruction in the same
3798 function.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3799
3800<h5>Semantics:</h5>
3801<p>The '<tt>resume</tt>' instruction resumes propagation of an existing
3802 (in-flight) exception whose unwinding was interrupted with
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3803 a <a href="#i_landingpad"><tt>landingpad</tt></a> instruction.</p>
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3804
3805<h5>Example:</h5>
3806<pre>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]3807 resume { i8*, i32 } %exn
Bill Wendlingdccc03b2011-07-31 06:30:59 +0000[diff] [blame]3808</pre>
3809
3810</div>
3811
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3812<!-- _______________________________________________________________________ -->
3813
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3814<h4>
3815 <a name="i_unreachable">'<tt>unreachable</tt>' Instruction</a>
3816</h4>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3817
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3818<div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3819
3820<h5>Syntax:</h5>
3821<pre>
3822 unreachable
3823</pre>
3824
3825<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3826<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3827 instruction is used to inform the optimizer that a particular portion of the
3828 code is not reachable. This can be used to indicate that the code after a
3829 no-return function cannot be reached, and other facts.</p>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3830
3831<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3832<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3833
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]3834</div>
3835
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3836</div>
3837
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3838<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3839<h3>
3840 <a name="binaryops">Binary Operations</a>
3841</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3842
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3843<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3844
3845<p>Binary operators are used to do most of the computation in a program. They
3846 require two operands of the same type, execute an operation on them, and
3847 produce a single value. The operands might represent multiple data, as is
3848 the case with the <a href="#t_vector">vector</a> data type. The result value
3849 has the same type as its operands.</p>
3850
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3851<p>There are several different binary operators:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3852
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3853<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3854<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3855 <a name="i_add">'<tt>add</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3856</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3857
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3858<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3859
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3860<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3861<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3862 &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]3863 &lt;result&gt; = add nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3864 &lt;result&gt; = add nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3865 &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]3866</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3867
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3868<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3869<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3870
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3871<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3872<p>The two arguments to the '<tt>add</tt>' instruction must
3873 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3874 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3875
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3876<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3877<p>The value produced is the integer sum of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3878
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3879<p>If the sum has unsigned overflow, the result returned is the mathematical
3880 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]3881
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3882<p>Because LLVM integers use a two's complement representation, this instruction
3883 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3884
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3885<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3886 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3887 <tt>nsw</tt> keywords are present, the result value of the <tt>add</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3888 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3889 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3890
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3891<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3892<pre>
3893 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3894</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3895
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3896</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3897
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3898<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3899<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3900 <a name="i_fadd">'<tt>fadd</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3901</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3902
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3903<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3904
3905<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3906<pre>
3907 &lt;result&gt; = fadd &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3908</pre>
3909
3910<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3911<p>The '<tt>fadd</tt>' instruction returns the sum of its two operands.</p>
3912
3913<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3914<p>The two arguments to the '<tt>fadd</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3915 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
3916 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3917
3918<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3919<p>The value produced is the floating point sum of the two operands.</p>
3920
3921<h5>Example:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3922<pre>
3923 &lt;result&gt; = fadd float 4.0, %var <i>; yields {float}:result = 4.0 + %var</i>
3924</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3925
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3926</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3927
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3928<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3929<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3930 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3931</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3932
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3933<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3934
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3935<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3936<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3937 &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]3938 &lt;result&gt; = sub nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3939 &lt;result&gt; = sub nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3940 &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]3941</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3942
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3943<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3944<p>The '<tt>sub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3945 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3946
3947<p>Note that the '<tt>sub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3948 '<tt>neg</tt>' instruction present in most other intermediate
3949 representations.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3950
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3951<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3952<p>The two arguments to the '<tt>sub</tt>' instruction must
3953 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
3954 integer values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3955
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3956<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3957<p>The value produced is the integer difference of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3958
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3959<p>If the difference has unsigned overflow, the result returned is the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3960 mathematical result modulo 2<sup>n</sup>, where n is the bit width of the
3961 result.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3962
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3963<p>Because LLVM integers use a two's complement representation, this instruction
3964 is appropriate for both signed and unsigned integers.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3965
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]3966<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
3967 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
3968 <tt>nsw</tt> keywords are present, the result value of the <tt>sub</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]3969 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]3970 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]3971
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3972<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]3973<pre>
3974 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3975 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3976</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3977
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3978</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]3979
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3980<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3981<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3982 <a name="i_fsub">'<tt>fsub</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]3983</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3984
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]3985<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3986
3987<h5>Syntax:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3988<pre>
3989 &lt;result&gt; = fsub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
3990</pre>
3991
3992<h5>Overview:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3993<p>The '<tt>fsub</tt>' instruction returns the difference of its two
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3994 operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3995
3996<p>Note that the '<tt>fsub</tt>' instruction is used to represent the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]3997 '<tt>fneg</tt>' instruction present in most other intermediate
3998 representations.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]3999
4000<h5>Arguments:</h5>
Bill Wendlingd9fe2982009-07-20 02:32:41 +0000[diff] [blame]4001<p>The two arguments to the '<tt>fsub</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4002 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4003 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4004
4005<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4006<p>The value produced is the floating point difference of the two operands.</p>
4007
4008<h5>Example:</h5>
4009<pre>
4010 &lt;result&gt; = fsub float 4.0, %var <i>; yields {float}:result = 4.0 - %var</i>
4011 &lt;result&gt; = fsub float -0.0, %val <i>; yields {float}:result = -%var</i>
4012</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4013
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4014</div>
4015
4016<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4017<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4018 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4019</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4020
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4021<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4022
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4023<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4024<pre>
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]4025 &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]4026 &lt;result&gt; = mul nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4027 &lt;result&gt; = mul nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4028 &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]4029</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4030
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4031<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4032<p>The '<tt>mul</tt>' instruction returns the product of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4033
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4034<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4035<p>The two arguments to the '<tt>mul</tt>' instruction must
4036 be <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
4037 integer values. Both arguments must have identical types.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4038
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4039<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4040<p>The value produced is the integer product of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4041
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4042<p>If the result of the multiplication has unsigned overflow, the result
4043 returned is the mathematical result modulo 2<sup>n</sup>, where n is the bit
4044 width of the result.</p>
4045
4046<p>Because LLVM integers use a two's complement representation, and the result
4047 is the same width as the operands, this instruction returns the correct
4048 result for both signed and unsigned integers. If a full product
4049 (e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands should
4050 be sign-extended or zero-extended as appropriate to the width of the full
4051 product.</p>
4052
Dan Gohman08d012e2009-07-22 22:44:56 +0000[diff] [blame]4053<p><tt>nuw</tt> and <tt>nsw</tt> stand for &quot;No Unsigned Wrap&quot;
4054 and &quot;No Signed Wrap&quot;, respectively. If the <tt>nuw</tt> and/or
4055 <tt>nsw</tt> keywords are present, the result value of the <tt>mul</tt>
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4056 is a <a href="#poisonvalues">poison value</a> if unsigned and/or signed overflow,
Dan Gohmanfff6c532010-04-22 23:14:21 +0000[diff] [blame]4057 respectively, occurs.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4058
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4059<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4060<pre>
4061 &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4062</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4063
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4064</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4065
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4066<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4067<h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4068 <a name="i_fmul">'<tt>fmul</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4069</h4>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4070
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4071<div>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4072
4073<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4074<pre>
4075 &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]4076</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4077
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4078<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4079<p>The '<tt>fmul</tt>' instruction returns the product of its two operands.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4080
4081<h5>Arguments:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4082<p>The two arguments to the '<tt>fmul</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4083 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4084 floating point values. Both arguments must have identical types.</p>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4085
4086<h5>Semantics:</h5>
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4087<p>The value produced is the floating point product of the two operands.</p>
4088
4089<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4090<pre>
4091 &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]4092</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4093
Dan Gohmanae3a0be2009-06-04 22:49:04 +0000[diff] [blame]4094</div>
4095
4096<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4097<h4>
4098 <a name="i_udiv">'<tt>udiv</tt>' Instruction</a>
4099</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4100
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4101<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4102
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4103<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4104<pre>
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]4105 &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4106 &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]4107</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4108
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4109<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4110<p>The '<tt>udiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4111
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4112<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4113<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4114 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4115 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4116
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4117<h5>Semantics:</h5>
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4118<p>The value produced is the unsigned integer quotient of the two operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4119
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4120<p>Note that unsigned integer division and signed integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4121 operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
4122
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4123<p>Division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4124
Chris Lattner35bda892011-02-06 21:44:57 +0000[diff] [blame]4125<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4126 <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]4127 multiple of %op2 (as such, "((a udiv exact b) mul b) == a").</p>
4128
4129
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4130<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4131<pre>
4132 &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4133</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4134
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4135</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4136
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4137<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4138<h4>
4139 <a name="i_sdiv">'<tt>sdiv</tt>' Instruction</a>
4140</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4141
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4142<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4143
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4144<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4145<pre>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4146 &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]4147 &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]4148</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4149
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4150<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4151<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4152
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4153<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4154<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4155 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4156 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4157
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4158<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4159<p>The value produced is the signed integer quotient of the two operands rounded
4160 towards zero.</p>
4161
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4162<p>Note that signed integer division and unsigned integer division are distinct
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4163 operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
4164
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4165<p>Division by zero leads to undefined behavior. Overflow also leads to
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4166 undefined behavior; this is a rare case, but can occur, for example, by doing
4167 a 32-bit division of -2147483648 by -1.</p>
4168
Dan Gohman9c5beed2009-07-22 00:04:19 +0000[diff] [blame]4169<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4170 <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]4171 be rounded.</p>
Dan Gohmancbb38f22009-07-20 22:41:19 +0000[diff] [blame]4172
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4173<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4174<pre>
4175 &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4176</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4177
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4178</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4179
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4180<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4181<h4>
4182 <a name="i_fdiv">'<tt>fdiv</tt>' Instruction</a>
4183</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4184
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4185<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4186
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4187<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4188<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4189 &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]4190</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4191
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4192<h5>Overview:</h5>
4193<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4194
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4195<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4196<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4197 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4198 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4199
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4200<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]4201<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4202
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4203<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4204<pre>
4205 &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]4206</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4207
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4208</div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4209
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4210<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4211<h4>
4212 <a name="i_urem">'<tt>urem</tt>' Instruction</a>
4213</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4214
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4215<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4216
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4217<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4218<pre>
4219 &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]4220</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4221
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4222<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4223<p>The '<tt>urem</tt>' instruction returns the remainder from the unsigned
4224 division of its two arguments.</p>
4225
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4226<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4227<p>The two arguments to the '<tt>urem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4228 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4229 values. Both arguments must have identical types.</p>
4230
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4231<h5>Semantics:</h5>
4232<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4233 This instruction always performs an unsigned division to get the
4234 remainder.</p>
4235
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4236<p>Note that unsigned integer remainder and signed integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4237 distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
4238
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4239<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4240
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4241<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4242<pre>
4243 &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4244</pre>
4245
4246</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4247
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4248<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4249<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4250 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4251</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4252
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4253<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4254
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4255<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4256<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4257 &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]4258</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4259
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4260<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4261<p>The '<tt>srem</tt>' instruction returns the remainder from the signed
4262 division of its two operands. This instruction can also take
4263 <a href="#t_vector">vector</a> versions of the values in which case the
4264 elements must be integers.</p>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]4265
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4266<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4267<p>The two arguments to the '<tt>srem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4268 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4269 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4270
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4271<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4272<p>This instruction returns the <i>remainder</i> of a division (where the result
Duncan Sandsdea3a5e2011-03-07 09:12:24 +0000[diff] [blame]4273 is either zero or has the same sign as the dividend, <tt>op1</tt>), not the
4274 <i>modulo</i> operator (where the result is either zero or has the same sign
4275 as the divisor, <tt>op2</tt>) of a value.
4276 For more information about the difference,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4277 see <a href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
4278 Math Forum</a>. For a table of how this is implemented in various languages,
4279 please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
4280 Wikipedia: modulo operation</a>.</p>
4281
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4282<p>Note that signed integer remainder and unsigned integer remainder are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4283 distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
4284
Chris Lattner5ec89832008-01-28 00:36:27 +0000[diff] [blame]4285<p>Taking the remainder of a division by zero leads to undefined behavior.
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4286 Overflow also leads to undefined behavior; this is a rare case, but can
4287 occur, for example, by taking the remainder of a 32-bit division of
4288 -2147483648 by -1. (The remainder doesn't actually overflow, but this rule
4289 lets srem be implemented using instructions that return both the result of
4290 the division and the remainder.)</p>
4291
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4292<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4293<pre>
4294 &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4295</pre>
4296
4297</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4298
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4299<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4300<h4>
4301 <a name="i_frem">'<tt>frem</tt>' Instruction</a>
4302</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4303
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4304<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4305
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4306<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4307<pre>
4308 &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]4309</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4310
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4311<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4312<p>The '<tt>frem</tt>' instruction returns the remainder from the division of
4313 its two operands.</p>
4314
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4315<h5>Arguments:</h5>
4316<p>The two arguments to the '<tt>frem</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4317 <a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a> of
4318 floating point values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4319
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4320<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4321<p>This instruction returns the <i>remainder</i> of a division. The remainder
4322 has the same sign as the dividend.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4323
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]4324<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4325<pre>
4326 &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]4327</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4328
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4329</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]4330
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4331</div>
4332
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4333<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4334<h3>
4335 <a name="bitwiseops">Bitwise Binary Operations</a>
4336</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4337
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4338<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4339
4340<p>Bitwise binary operators are used to do various forms of bit-twiddling in a
4341 program. They are generally very efficient instructions and can commonly be
4342 strength reduced from other instructions. They require two operands of the
4343 same type, execute an operation on them, and produce a single value. The
4344 resulting value is the same type as its operands.</p>
4345
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4346<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4347<h4>
4348 <a name="i_shl">'<tt>shl</tt>' Instruction</a>
4349</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4350
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4351<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4352
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4353<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4354<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4355 &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4356 &lt;result&gt; = shl nuw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4357 &lt;result&gt; = shl nsw &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4358 &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]4359</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4360
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4361<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4362<p>The '<tt>shl</tt>' instruction returns the first operand shifted to the left
4363 a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4364
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4365<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4366<p>Both arguments to the '<tt>shl</tt>' instruction must be the
4367 same <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of
4368 integer type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4369
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4370<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4371<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod
4372 2<sup>n</sup>, where <tt>n</tt> is the width of the result. If <tt>op2</tt>
4373 is (statically or dynamically) negative or equal to or larger than the number
4374 of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4375 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4376 shift amount in <tt>op2</tt>.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4377
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4378<p>If the <tt>nuw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4379 <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]4380 the <tt>nsw</tt> keyword is present, then the shift produces a
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4381 <a href="#poisonvalues">poison value</a> if it shifts out any bits that disagree
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4382 with the resultant sign bit. As such, NUW/NSW have the same semantics as
4383 they would if the shift were expressed as a mul instruction with the same
4384 nsw/nuw bits in (mul %op1, (shl 1, %op2)).</p>
4385
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4386<h5>Example:</h5>
4387<pre>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4388 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
4389 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
4390 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4391 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4392 &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]4393</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4394
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4395</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4396
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4397<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4398<h4>
4399 <a name="i_lshr">'<tt>lshr</tt>' Instruction</a>
4400</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4401
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4402<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4403
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4404<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4405<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4406 &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4407 &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]4408</pre>
4409
4410<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4411<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
4412 operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4413
4414<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4415<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4416 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4417 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4418
4419<h5>Semantics:</h5>
4420<p>This instruction always performs a logical shift right operation. The most
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4421 significant bits of the result will be filled with zero bits after the shift.
4422 If <tt>op2</tt> is (statically or dynamically) equal to or larger than the
4423 number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
4424 vectors, each vector element of <tt>op1</tt> is shifted by the corresponding
4425 shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4426
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4427<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4428 <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]4429 shifted out are non-zero.</p>
4430
4431
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4432<h5>Example:</h5>
4433<pre>
4434 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
4435 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
4436 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
4437 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4438 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4439 &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]4440</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4441
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4442</div>
4443
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]4444<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4445<h4>
4446 <a name="i_ashr">'<tt>ashr</tt>' Instruction</a>
4447</h4>
4448
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4449<div>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4450
4451<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4452<pre>
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4453 &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4454 &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]4455</pre>
4456
4457<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4458<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
4459 operand shifted to the right a specified number of bits with sign
4460 extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4461
4462<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4463<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4464 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4465 type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4466
4467<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4468<p>This instruction always performs an arithmetic shift right operation, The
4469 most significant bits of the result will be filled with the sign bit
4470 of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
4471 larger than the number of bits in <tt>op1</tt>, the result is undefined. If
4472 the arguments are vectors, each vector element of <tt>op1</tt> is shifted by
4473 the corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4474
Chris Lattnerf067d582011-02-07 16:40:21 +0000[diff] [blame]4475<p>If the <tt>exact</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]4476 <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]4477 shifted out are non-zero.</p>
4478
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4479<h5>Example:</h5>
4480<pre>
4481 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
4482 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
4483 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
4484 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]4485 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wange9f10152008-12-09 05:46:39 +0000[diff] [blame]4486 &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]4487</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4488
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]4489</div>
4490
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4491<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4492<h4>
4493 <a name="i_and">'<tt>and</tt>' Instruction</a>
4494</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4495
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4496<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4497
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4498<h5>Syntax:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4499<pre>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]4500 &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]4501</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4502
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4503<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4504<p>The '<tt>and</tt>' instruction returns the bitwise logical and of its two
4505 operands.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4506
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4507<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4508<p>The two arguments to the '<tt>and</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4509 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4510 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4511
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4512<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4513<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4514
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4515<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4516 <tbody>
4517 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4518 <th>In0</th>
4519 <th>In1</th>
4520 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4521 </tr>
4522 <tr>
4523 <td>0</td>
4524 <td>0</td>
4525 <td>0</td>
4526 </tr>
4527 <tr>
4528 <td>0</td>
4529 <td>1</td>
4530 <td>0</td>
4531 </tr>
4532 <tr>
4533 <td>1</td>
4534 <td>0</td>
4535 <td>0</td>
4536 </tr>
4537 <tr>
4538 <td>1</td>
4539 <td>1</td>
4540 <td>1</td>
4541 </tr>
4542 </tbody>
4543</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4544
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4545<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4546<pre>
4547 &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]4548 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
4549 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4550</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4551</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4552<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4553<h4>
4554 <a name="i_or">'<tt>or</tt>' Instruction</a>
4555</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4556
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4557<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4558
4559<h5>Syntax:</h5>
4560<pre>
4561 &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
4562</pre>
4563
4564<h5>Overview:</h5>
4565<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive or of its
4566 two operands.</p>
4567
4568<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4569<p>The two arguments to the '<tt>or</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4570 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4571 values. Both arguments must have identical types.</p>
4572
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4573<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4574<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4575
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4576<table border="1" cellspacing="0" cellpadding="4">
4577 <tbody>
4578 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4579 <th>In0</th>
4580 <th>In1</th>
4581 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4582 </tr>
4583 <tr>
4584 <td>0</td>
4585 <td>0</td>
4586 <td>0</td>
4587 </tr>
4588 <tr>
4589 <td>0</td>
4590 <td>1</td>
4591 <td>1</td>
4592 </tr>
4593 <tr>
4594 <td>1</td>
4595 <td>0</td>
4596 <td>1</td>
4597 </tr>
4598 <tr>
4599 <td>1</td>
4600 <td>1</td>
4601 <td>1</td>
4602 </tr>
4603 </tbody>
4604</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4605
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4606<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4607<pre>
4608 &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4609 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
4610 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4611</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4612
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4613</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4614
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4615<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4616<h4>
4617 <a name="i_xor">'<tt>xor</tt>' Instruction</a>
4618</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4619
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4620<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4621
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4622<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4623<pre>
4624 &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]4625</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4626
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4627<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4628<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive or of
4629 its two operands. The <tt>xor</tt> is used to implement the "one's
4630 complement" operation, which is the "~" operator in C.</p>
4631
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4632<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4633<p>The two arguments to the '<tt>xor</tt>' instruction must be
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4634 <a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
4635 values. Both arguments must have identical types.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]4636
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4637<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4638<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4639
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4640<table border="1" cellspacing="0" cellpadding="4">
4641 <tbody>
4642 <tr>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]4643 <th>In0</th>
4644 <th>In1</th>
4645 <th>Out</th>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]4646 </tr>
4647 <tr>
4648 <td>0</td>
4649 <td>0</td>
4650 <td>0</td>
4651 </tr>
4652 <tr>
4653 <td>0</td>
4654 <td>1</td>
4655 <td>1</td>
4656 </tr>
4657 <tr>
4658 <td>1</td>
4659 <td>0</td>
4660 <td>1</td>
4661 </tr>
4662 <tr>
4663 <td>1</td>
4664 <td>1</td>
4665 <td>0</td>
4666 </tr>
4667 </tbody>
4668</table>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4669
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4670<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4671<pre>
4672 &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4673 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
4674 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
4675 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4676</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4677
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4678</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4679
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4680</div>
4681
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4682<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4683<h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4684 <a name="vectorops">Vector Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4685</h3>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4686
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4687<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4688
4689<p>LLVM supports several instructions to represent vector operations in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4690 target-independent manner. These instructions cover the element-access and
4691 vector-specific operations needed to process vectors effectively. While LLVM
4692 does directly support these vector operations, many sophisticated algorithms
4693 will want to use target-specific intrinsics to take full advantage of a
4694 specific target.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4695
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4696<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4697<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4698 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4699</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4700
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4701<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4702
4703<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4704<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4705 &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]4706</pre>
4707
4708<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4709<p>The '<tt>extractelement</tt>' instruction extracts a single scalar element
4710 from a vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4711
4712
4713<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4714<p>The first operand of an '<tt>extractelement</tt>' instruction is a value
4715 of <a href="#t_vector">vector</a> type. The second operand is an index
4716 indicating the position from which to extract the element. The index may be
4717 a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4718
4719<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4720<p>The result is a scalar of the same type as the element type of
4721 <tt>val</tt>. Its value is the value at position <tt>idx</tt> of
4722 <tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4723 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4724
4725<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4726<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4727 &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]4728</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4729
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4730</div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4731
4732<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4733<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4734 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4735</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4736
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4737<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4738
4739<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4740<pre>
Dan Gohmanf3480b92008-05-12 23:38:42 +0000[diff] [blame]4741 &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]4742</pre>
4743
4744<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4745<p>The '<tt>insertelement</tt>' instruction inserts a scalar element into a
4746 vector at a specified index.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4747
4748<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4749<p>The first operand of an '<tt>insertelement</tt>' instruction is a value
4750 of <a href="#t_vector">vector</a> type. The second operand is a scalar value
4751 whose type must equal the element type of the first operand. The third
4752 operand is an index indicating the position at which to insert the value.
4753 The index may be a variable.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4754
4755<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4756<p>The result is a vector of the same type as <tt>val</tt>. Its element values
4757 are those of <tt>val</tt> except at position <tt>idx</tt>, where it gets the
4758 value <tt>elt</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
4759 results are undefined.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4760
4761<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4762<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4763 &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]4764</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4765
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4766</div>
4767
4768<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4769<h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4770 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4771</h4>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4772
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4773<div>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4774
4775<h5>Syntax:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4776<pre>
Mon P Wangaeb06d22008-11-10 04:46:22 +0000[diff] [blame]4777 &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]4778</pre>
4779
4780<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4781<p>The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
4782 from two input vectors, returning a vector with the same element type as the
4783 input and length that is the same as the shuffle mask.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4784
4785<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4786<p>The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
Duncan Sandsb5a1bf42012-06-14 14:58:28 +0000[diff] [blame]4787 with the same type. The third argument is a shuffle mask whose
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4788 element type is always 'i32'. The result of the instruction is a vector
4789 whose length is the same as the shuffle mask and whose element type is the
4790 same as the element type of the first two operands.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4791
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4792<p>The shuffle mask operand is required to be a constant vector with either
4793 constant integer or undef values.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4794
4795<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4796<p>The elements of the two input vectors are numbered from left to right across
4797 both of the vectors. The shuffle mask operand specifies, for each element of
4798 the result vector, which element of the two input vectors the result element
4799 gets. The element selector may be undef (meaning "don't care") and the
4800 second operand may be undef if performing a shuffle from only one vector.</p>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4801
4802<h5>Example:</h5>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4803<pre>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]4804 &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]4805 &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]4806 &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]4807 &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]4808 &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]4809 &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]4810 &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]4811 &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]4812</pre>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4813
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4814</div>
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]4815
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4816</div>
4817
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]4818<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4819<h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4820 <a name="aggregateops">Aggregate Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4821</h3>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4822
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4823<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4824
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4825<p>LLVM supports several instructions for working with
4826 <a href="#t_aggregate">aggregate</a> values.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4827
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4828<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4829<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4830 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4831</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4832
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4833<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4834
4835<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4836<pre>
4837 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
4838</pre>
4839
4840<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4841<p>The '<tt>extractvalue</tt>' instruction extracts the value of a member field
4842 from an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4843
4844<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4845<p>The first operand of an '<tt>extractvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4846 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4847 <a href="#t_array">array</a> type. The operands are constant indices to
4848 specify which value to extract in a similar manner as indices in a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4849 '<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.</p>
Frits van Bommel13242892010-12-05 20:54:38 +0000[diff] [blame]4850 <p>The major differences to <tt>getelementptr</tt> indexing are:</p>
4851 <ul>
4852 <li>Since the value being indexed is not a pointer, the first index is
4853 omitted and assumed to be zero.</li>
4854 <li>At least one index must be specified.</li>
4855 <li>Not only struct indices but also array indices must be in
4856 bounds.</li>
4857 </ul>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4858
4859<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4860<p>The result is the value at the position in the aggregate specified by the
4861 index operands.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4862
4863<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4864<pre>
Gabor Greifa5b6f452009-10-28 13:14:50 +0000[diff] [blame]4865 &lt;result&gt; = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4866</pre>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4867
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4868</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4869
4870<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4871<h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4872 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4873</h4>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4874
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4875<div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4876
4877<h5>Syntax:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4878<pre>
Bill Wendling194229e2011-07-26 20:42:28 +0000[diff] [blame]4879 &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]4880</pre>
4881
4882<h5>Overview:</h5>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4883<p>The '<tt>insertvalue</tt>' instruction inserts a value into a member field
4884 in an <a href="#t_aggregate">aggregate</a> value.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4885
4886<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4887<p>The first operand of an '<tt>insertvalue</tt>' instruction is a value
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]4888 of <a href="#t_struct">struct</a> or
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]4889 <a href="#t_array">array</a> type. The second operand is a first-class
4890 value to insert. The following operands are constant indices indicating
4891 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]4892 '<tt><a href="#i_extractvalue">extractvalue</a></tt>' instruction. The
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4893 value to insert must have the same type as the value identified by the
4894 indices.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4895
4896<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4897<p>The result is an aggregate of the same type as <tt>val</tt>. Its value is
4898 that of <tt>val</tt> except that the value at the position specified by the
4899 indices is that of <tt>elt</tt>.</p>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4900
4901<h5>Example:</h5>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4902<pre>
Chris Lattner8645d1a2011-05-22 07:18:08 +0000[diff] [blame]4903 %agg1 = insertvalue {i32, float} undef, i32 1, 0 <i>; yields {i32 1, float undef}</i>
4904 %agg2 = insertvalue {i32, float} %agg1, float %val, 1 <i>; yields {i32 1, float %val}</i>
4905 %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]4906</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4907
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4908</div>
4909
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4910</div>
Dan Gohmana334d5f2008-05-12 23:51:09 +0000[diff] [blame]4911
4912<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4913<h3>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]4914 <a name="memoryops">Memory Access and Addressing Operations</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4915</h3>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4916
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4917<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4918
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4919<p>A key design point of an SSA-based representation is how it represents
4920 memory. In LLVM, no memory locations are in SSA form, which makes things
Victor Hernandez2fee2942009-10-26 23:44:29 +0000[diff] [blame]4921 very simple. This section describes how to read, write, and allocate
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4922 memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4923
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4924<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4925<h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4926 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4927</h4>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4928
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4929<div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4930
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4931<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4932<pre>
Dan Gohmanf75a7d32010-05-28 01:14:11 +0000[diff] [blame]4933 &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]4934</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4935
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4936<h5>Overview:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]4937<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4938 currently executing function, to be automatically released when this function
4939 returns to its caller. The object is always allocated in the generic address
4940 space (address space zero).</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4941
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4942<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4943<p>The '<tt>alloca</tt>' instruction
4944 allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt> bytes of memory on the
4945 runtime stack, returning a pointer of the appropriate type to the program.
4946 If "NumElements" is specified, it is the number of elements allocated,
4947 otherwise "NumElements" is defaulted to be one. If a constant alignment is
4948 specified, the value result of the allocation is guaranteed to be aligned to
4949 at least that boundary. If not specified, or if zero, the target can choose
4950 to align the allocation on any convenient boundary compatible with the
4951 type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4952
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4953<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4954
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4955<h5>Semantics:</h5>
Bill Wendling871eb0a2009-05-08 20:49:29 +0000[diff] [blame]4956<p>Memory is allocated; a pointer is returned. The operation is undefined if
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4957 there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
4958 memory is automatically released when the function returns. The
4959 '<tt>alloca</tt>' instruction is commonly used to represent automatic
4960 variables that must have an address available. When the function returns
4961 (either with the <tt><a href="#i_ret">ret</a></tt>
Bill Wendling7b9e5392012-02-06 21:57:33 +0000[diff] [blame]4962 or <tt><a href="#i_resume">resume</a></tt> instructions), the memory is
Nick Lewycky84a1d232012-02-29 08:26:44 +0000[diff] [blame]4963 reclaimed. Allocating zero bytes is legal, but the result is undefined.
4964 The order in which memory is allocated (ie., which way the stack grows) is
Nick Lewycky75d05e62012-03-18 09:35:50 +0000[diff] [blame]4965 not specified.</p>
Nick Lewycky84a1d232012-02-29 08:26:44 +0000[diff] [blame]4966
4967<p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4968
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4969<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4970<pre>
Dan Gohman81e21672009-01-04 23:49:44 +0000[diff] [blame]4971 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
4972 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
4973 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
4974 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4975</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4976
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4977</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]4978
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]4979<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]4980<h4>
4981 <a name="i_load">'<tt>load</tt>' Instruction</a>
4982</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4983
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]4984<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4985
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4986<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4987<pre>
Pete Cooperf95acc62012-02-10 18:13:54 +0000[diff] [blame]4988 &lt;result&gt; = load [volatile] &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;][, !nontemporal !&lt;index&gt;][, !invariant.load !&lt;index&gt;]
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]4989 &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]4990 !&lt;index&gt; = !{ i32 1 }
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4991</pre>
4992
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4993<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]4994<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4995
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]4996<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]4997<p>The argument to the '<tt>load</tt>' instruction specifies the memory address
4998 from which to load. The pointer must point to
4999 a <a href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
5000 marked as <tt>volatile</tt>, then the optimizer is not allowed to modify the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]5001 number or order of execution of this <tt>load</tt> with other <a
5002 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5003
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]5004<p>If the <code>load</code> is marked as <code>atomic</code>, it takes an extra
5005 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
5006 argument. The <code>release</code> and <code>acq_rel</code> orderings are
5007 not valid on <code>load</code> instructions. Atomic loads produce <a
5008 href="#memorymodel">defined</a> results when they may see multiple atomic
5009 stores. The type of the pointee must be an integer type whose bit width
5010 is a power of two greater than or equal to eight and less than or equal
5011 to a target-specific size limit. <code>align</code> must be explicitly
5012 specified on atomic loads, and the load has undefined behavior if the
5013 alignment is not set to a value which is at least the size in bytes of
5014 the pointee. <code>!nontemporal</code> does not have any defined semantics
5015 for atomic loads.</p>
5016
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]5017<p>The optional constant <tt>align</tt> argument specifies the alignment of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5018 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]5019 omitted <tt>align</tt> argument means that the operation has the preferential
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5020 alignment for the target. It is the responsibility of the code emitter to
5021 ensure that the alignment information is correct. Overestimating the
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]5022 alignment results in undefined behavior. Underestimating the alignment may
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5023 produce less efficient code. An alignment of 1 is always safe.</p>
5024
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]5025<p>The optional <tt>!nontemporal</tt> metadata must reference a single
5026 metatadata name &lt;index&gt; corresponding to a metadata node with
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5027 one <tt>i32</tt> entry of value 1. The existence of
Bill Wendling7c78dbb2010-02-25 21:23:24 +0000[diff] [blame]5028 the <tt>!nontemporal</tt> metatadata on the instruction tells the optimizer
5029 and code generator that this load is not expected to be reused in the cache.
5030 The code generator may select special instructions to save cache bandwidth,
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5031 such as the <tt>MOVNT</tt> instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5032
Pete Cooperf95acc62012-02-10 18:13:54 +0000[diff] [blame]5033<p>The optional <tt>!invariant.load</tt> metadata must reference a single
5034 metatadata name &lt;index&gt; corresponding to a metadata node with no
5035 entries. The existence of the <tt>!invariant.load</tt> metatadata on the
5036 instruction tells the optimizer and code generator that this load address
5037 points to memory which does not change value during program execution.
5038 The optimizer may then move this load around, for example, by hoisting it
5039 out of loops using loop invariant code motion.</p>
5040
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5041<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5042<p>The location of memory pointed to is loaded. If the value being loaded is of
5043 scalar type then the number of bytes read does not exceed the minimum number
5044 of bytes needed to hold all bits of the type. For example, loading an
5045 <tt>i24</tt> reads at most three bytes. When loading a value of a type like
5046 <tt>i20</tt> with a size that is not an integral number of bytes, the result
5047 is undefined if the value was not originally written using a store of the
5048 same type.</p>
5049
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5050<h5>Examples:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5051<pre>
5052 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
5053 <a href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5054 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5055</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5056
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5057</div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5058
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5059<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5060<h4>
5061 <a name="i_store">'<tt>store</tt>' Instruction</a>
5062</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5063
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5064<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5065
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5066<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5067<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5068 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>
5069 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]5070</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5071
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5072<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5073<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5074
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5075<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5076<p>There are two arguments to the '<tt>store</tt>' instruction: a value to store
5077 and an address at which to store it. The type of the
5078 '<tt>&lt;pointer&gt;</tt>' operand must be a pointer to
5079 the <a href="#t_firstclass">first class</a> type of the
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]5080 '<tt>&lt;value&gt;</tt>' operand. If the <tt>store</tt> is marked as
5081 <tt>volatile</tt>, then the optimizer is not allowed to modify the number or
5082 order of execution of this <tt>store</tt> with other <a
5083 href="#volatile">volatile operations</a>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5084
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]5085<p>If the <code>store</code> is marked as <code>atomic</code>, it takes an extra
5086 <a href="#ordering">ordering</a> and optional <code>singlethread</code>
5087 argument. The <code>acquire</code> and <code>acq_rel</code> orderings aren't
5088 valid on <code>store</code> instructions. Atomic loads produce <a
5089 href="#memorymodel">defined</a> results when they may see multiple atomic
5090 stores. The type of the pointee must be an integer type whose bit width
5091 is a power of two greater than or equal to eight and less than or equal
5092 to a target-specific size limit. <code>align</code> must be explicitly
5093 specified on atomic stores, and the store has undefined behavior if the
5094 alignment is not set to a value which is at least the size in bytes of
5095 the pointee. <code>!nontemporal</code> does not have any defined semantics
5096 for atomic stores.</p>
5097
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5098<p>The optional constant "align" argument specifies the alignment of the
5099 operation (that is, the alignment of the memory address). A value of 0 or an
5100 omitted "align" argument means that the operation has the preferential
5101 alignment for the target. It is the responsibility of the code emitter to
5102 ensure that the alignment information is correct. Overestimating the
5103 alignment results in an undefined behavior. Underestimating the alignment may
5104 produce less efficient code. An alignment of 1 is always safe.</p>
5105
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5106<p>The optional !nontemporal metadata must reference a single metatadata
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5107 name &lt;index&gt; corresponding to a metadata node with one i32 entry of
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5108 value 1. The existence of the !nontemporal metatadata on the
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5109 instruction tells the optimizer and code generator that this load is
5110 not expected to be reused in the cache. The code generator may
5111 select special instructions to save cache bandwidth, such as the
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]5112 MOVNT instruction on x86.</p>
David Greene8939b0d2010-02-16 20:50:18 +0000[diff] [blame]5113
5114
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]5115<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5116<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>' at the
5117 location specified by the '<tt>&lt;pointer&gt;</tt>' operand. If
5118 '<tt>&lt;value&gt;</tt>' is of scalar type then the number of bytes written
5119 does not exceed the minimum number of bytes needed to hold all bits of the
5120 type. For example, storing an <tt>i24</tt> writes at most three bytes. When
5121 writing a value of a type like <tt>i20</tt> with a size that is not an
5122 integral number of bytes, it is unspecified what happens to the extra bits
5123 that do not belong to the type, but they will typically be overwritten.</p>
5124
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5125<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5126<pre>
5127 %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]5128 store i32 3, i32* %ptr <i>; yields {void}</i>
5129 %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]5130</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5131
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5132</div>
5133
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]5134<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5135<h4>
5136<a name="i_fence">'<tt>fence</tt>' Instruction</a>
5137</h4>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5138
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5139<div>
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5140
5141<h5>Syntax:</h5>
5142<pre>
5143 fence [singlethread] &lt;ordering&gt; <i>; yields {void}</i>
5144</pre>
5145
5146<h5>Overview:</h5>
5147<p>The '<tt>fence</tt>' instruction is used to introduce happens-before edges
5148between operations.</p>
5149
5150<h5>Arguments:</h5> <p>'<code>fence</code>' instructions take an <a
5151href="#ordering">ordering</a> argument which defines what
5152<i>synchronizes-with</i> edges they add. They can only be given
5153<code>acquire</code>, <code>release</code>, <code>acq_rel</code>, and
5154<code>seq_cst</code> orderings.</p>
5155
5156<h5>Semantics:</h5>
5157<p>A fence <var>A</var> which has (at least) <code>release</code> ordering
5158semantics <i>synchronizes with</i> a fence <var>B</var> with (at least)
5159<code>acquire</code> ordering semantics if and only if there exist atomic
5160operations <var>X</var> and <var>Y</var>, both operating on some atomic object
5161<var>M</var>, such that <var>A</var> is sequenced before <var>X</var>,
5162<var>X</var> modifies <var>M</var> (either directly or through some side effect
5163of a sequence headed by <var>X</var>), <var>Y</var> is sequenced before
5164<var>B</var>, and <var>Y</var> observes <var>M</var>. This provides a
5165<i>happens-before</i> dependency between <var>A</var> and <var>B</var>. Rather
5166than an explicit <code>fence</code>, one (but not both) of the atomic operations
5167<var>X</var> or <var>Y</var> might provide a <code>release</code> or
5168<code>acquire</code> (resp.) ordering constraint and still
5169<i>synchronize-with</i> the explicit <code>fence</code> and establish the
5170<i>happens-before</i> edge.</p>
5171
5172<p>A <code>fence</code> which has <code>seq_cst</code> ordering, in addition to
5173having both <code>acquire</code> and <code>release</code> semantics specified
5174above, participates in the global program order of other <code>seq_cst</code>
5175operations and/or fences.</p>
5176
5177<p>The optional "<a href="#singlethread"><code>singlethread</code></a>" argument
5178specifies that the fence only synchronizes with other fences in the same
5179thread. (This is useful for interacting with signal handlers.)</p>
5180
Eli Friedman47f35132011-07-25 23:16:38 +0000[diff] [blame]5181<h5>Example:</h5>
5182<pre>
5183 fence acquire <i>; yields {void}</i>
5184 fence singlethread seq_cst <i>; yields {void}</i>
5185</pre>
5186
5187</div>
5188
5189<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5190<h4>
5191<a name="i_cmpxchg">'<tt>cmpxchg</tt>' Instruction</a>
5192</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5193
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5194<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5195
5196<h5>Syntax:</h5>
5197<pre>
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5198 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]5199</pre>
5200
5201<h5>Overview:</h5>
5202<p>The '<tt>cmpxchg</tt>' instruction is used to atomically modify memory.
5203It loads a value in memory and compares it to a given value. If they are
5204equal, it stores a new value into the memory.</p>
5205
5206<h5>Arguments:</h5>
5207<p>There are three arguments to the '<code>cmpxchg</code>' instruction: an
5208address to operate on, a value to compare to the value currently be at that
5209address, and a new value to place at that address if the compared values are
5210equal. The type of '<var>&lt;cmp&gt;</var>' must be an integer type whose
5211bit width is a power of two greater than or equal to eight and less than
5212or equal to a target-specific size limit. '<var>&lt;cmp&gt;</var>' and
5213'<var>&lt;new&gt;</var>' must have the same type, and the type of
5214'<var>&lt;pointer&gt;</var>' must be a pointer to that type. If the
5215<code>cmpxchg</code> is marked as <code>volatile</code>, then the
5216optimizer is not allowed to modify the number or order of execution
5217of this <code>cmpxchg</code> with other <a href="#volatile">volatile
5218operations</a>.</p>
5219
5220<!-- FIXME: Extend allowed types. -->
5221
5222<p>The <a href="#ordering"><var>ordering</var></a> argument specifies how this
5223<code>cmpxchg</code> synchronizes with other atomic operations.</p>
5224
5225<p>The optional "<code>singlethread</code>" argument declares that the
5226<code>cmpxchg</code> is only atomic with respect to code (usually signal
5227handlers) running in the same thread as the <code>cmpxchg</code>. Otherwise the
5228cmpxchg is atomic with respect to all other code in the system.</p>
5229
5230<p>The pointer passed into cmpxchg must have alignment greater than or equal to
5231the size in memory of the operand.
5232
5233<h5>Semantics:</h5>
5234<p>The contents of memory at the location specified by the
5235'<tt>&lt;pointer&gt;</tt>' operand is read and compared to
5236'<tt>&lt;cmp&gt;</tt>'; if the read value is the equal,
5237'<tt>&lt;new&gt;</tt>' is written. The original value at the location
5238is returned.
5239
5240<p>A successful <code>cmpxchg</code> is a read-modify-write instruction for the
5241purpose of identifying <a href="#release_sequence">release sequences</a>. A
5242failed <code>cmpxchg</code> is equivalent to an atomic load with an ordering
5243parameter determined by dropping any <code>release</code> part of the
5244<code>cmpxchg</code>'s ordering.</p>
5245
5246<!--
5247FIXME: Is compare_exchange_weak() necessary? (Consider after we've done
5248optimization work on ARM.)
5249
5250FIXME: Is a weaker ordering constraint on failure helpful in practice?
5251-->
5252
5253<h5>Example:</h5>
5254<pre>
5255entry:
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5256 %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]5257 <a href="#i_br">br</a> label %loop
5258
5259loop:
5260 %cmp = <a href="#i_phi">phi</a> i32 [ %orig, %entry ], [%old, %loop]
5261 %squared = <a href="#i_mul">mul</a> i32 %cmp, %cmp
Bill Wendling262396b2011-12-09 22:41:40 +0000[diff] [blame]5262 %old = cmpxchg i32* %ptr, i32 %cmp, i32 %squared <i>; yields {i32}</i>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5263 %success = <a href="#i_icmp">icmp</a> eq i32 %cmp, %old
5264 <a href="#i_br">br</a> i1 %success, label %done, label %loop
5265
5266done:
5267 ...
5268</pre>
5269
5270</div>
5271
5272<!-- _______________________________________________________________________ -->
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5273<h4>
5274<a name="i_atomicrmw">'<tt>atomicrmw</tt>' Instruction</a>
5275</h4>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5276
NAKAMURA Takumib5bb29c2011-08-12 06:17:17 +0000[diff] [blame]5277<div>
Eli Friedmanff030482011-07-28 21:48:00 +0000[diff] [blame]5278
5279<h5>Syntax:</h5>
5280<pre>
Eli Friedmanf03bb262011-08-12 22:50:01 +0000[diff] [blame]5281 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]5282</pre>
5283
5284<h5>Overview:</h5>
5285<p>The '<tt>atomicrmw</tt>' instruction is used to atomically modify memory.</p>
5286
5287<h5>Arguments:</h5>
5288<p>There are three arguments to the '<code>atomicrmw</code>' instruction: an
5289operation to apply, an address whose value to modify, an argument to the
5290operation. The operation must be one of the following keywords:</p>
5291<ul>
5292 <li>xchg</li>
5293 <li>add</li>
5294 <li>sub</li>
5295 <li>and</li>
5296 <li>nand</li>
5297 <li>or</li>
5298 <li>xor</li>
5299 <li>max</li>
5300 <li>min</li>
5301 <li>umax</li>
5302 <li>umin</li>
5303</ul>
5304
5305<p>The type of '<var>&lt;value&gt;</var>' must be an integer type whose
5306bit width is a power of two greater than or equal to eight and less than
5307or equal to a target-specific size limit. The type of the
5308'<code>&lt;pointer&gt;</code>' operand must be a pointer to that type.
5309If the <code>atomicrmw</code> is marked as <code>volatile</code>, then the
5310optimizer is not allowed to modify the number or order of execution of this
5311<code>atomicrmw</code> with other <a href="#volatile">volatile
5312 operations</a>.</p>
5313
5314<!-- FIXME: Extend allowed types. -->
5315
5316<h5>Semantics:</h5>
5317<p>The contents of memory at the location specified by the
5318'<tt>&lt;pointer&gt;</tt>' operand are atomically read, modified, and written
5319back. The original value at the location is returned. The modification is
5320specified by the <var>operation</var> argument:</p>
5321
5322<ul>
5323 <li>xchg: <code>*ptr = val</code></li>
5324 <li>add: <code>*ptr = *ptr + val</code></li>
5325 <li>sub: <code>*ptr = *ptr - val</code></li>
5326 <li>and: <code>*ptr = *ptr &amp; val</code></li>
5327 <li>nand: <code>*ptr = ~(*ptr &amp; val)</code></li>
5328 <li>or: <code>*ptr = *ptr | val</code></li>
5329 <li>xor: <code>*ptr = *ptr ^ val</code></li>
5330 <li>max: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using a signed comparison)</li>
5331 <li>min: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using a signed comparison)</li>
5332 <li>umax: <code>*ptr = *ptr &gt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5333 <li>umin: <code>*ptr = *ptr &lt; val ? *ptr : val</code> (using an unsigned comparison)</li>
5334</ul>
5335
5336<h5>Example:</h5>
5337<pre>
5338 %old = atomicrmw add i32* %ptr, i32 1 acquire <i>; yields {i32}</i>
5339</pre>
5340
5341</div>
5342
5343<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5344<h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5345 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5346</h4>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5347
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5348<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5349
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5350<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5351<pre>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5352 &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]5353 &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]5354 &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]5355</pre>
5356
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5357<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5358<p>The '<tt>getelementptr</tt>' instruction is used to get the address of a
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5359 subelement of an <a href="#t_aggregate">aggregate</a> data structure.
5360 It performs address calculation only and does not access memory.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5361
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5362<h5>Arguments:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5363<p>The first argument is always a pointer or a vector of pointers,
5364 and forms the basis of the
Chris Lattnerc8eef442009-07-29 06:44:13 +0000[diff] [blame]5365 calculation. The remaining arguments are indices that indicate which of the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5366 elements of the aggregate object are indexed. The interpretation of each
5367 index is dependent on the type being indexed into. The first index always
5368 indexes the pointer value given as the first argument, the second index
5369 indexes a value of the type pointed to (not necessarily the value directly
5370 pointed to, since the first index can be non-zero), etc. The first type
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5371 indexed into must be a pointer value, subsequent types can be arrays,
Chris Lattner61c70e92010-08-28 04:09:24 +0000[diff] [blame]5372 vectors, and structs. Note that subsequent types being indexed into
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5373 can never be pointers, since that would require loading the pointer before
5374 continuing calculation.</p>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5375
5376<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]5377 When indexing into a (optionally packed) structure, only <tt>i32</tt>
Chris Lattnerfdfeb692010-02-12 20:49:41 +0000[diff] [blame]5378 integer <b>constants</b> are allowed. When indexing into an array, pointer
5379 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]5380 constant. These integers are treated as signed values where relevant.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5381
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5382<p>For example, let's consider a C code fragment and how it gets compiled to
5383 LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5384
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5385<pre class="doc_code">
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5386struct RT {
5387 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5388 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5389 char C;
5390};
5391struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5392 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5393 double Y;
5394 struct RT Z;
5395};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5396
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]5397int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5398 return &amp;s[1].Z.B[5][13];
5399}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5400</pre>
5401
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5402<p>The LLVM code generated by Clang is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5403
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]5404<pre class="doc_code">
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5405%struct.RT = <a href="#namedtypes">type</a> { i8, [10 x [20 x i32]], i8 }
5406%struct.ST = <a href="#namedtypes">type</a> { i32, double, %struct.RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5407
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5408define i32* @foo(%struct.ST* %s) nounwind uwtable readnone optsize ssp {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5409entry:
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5410 %arrayidx = getelementptr inbounds %struct.ST* %s, i64 1, i32 2, i32 1, i64 5, i64 13
5411 ret i32* %arrayidx
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]5412}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5413</pre>
5414
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5415<h5>Semantics:</h5>
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5416<p>In the example above, the first index is indexing into the
5417 '<tt>%struct.ST*</tt>' type, which is a pointer, yielding a
5418 '<tt>%struct.ST</tt>' = '<tt>{ i32, double, %struct.RT }</tt>' type, a
5419 structure. The second index indexes into the third element of the structure,
5420 yielding a '<tt>%struct.RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]], i8 }</tt>'
5421 type, another structure. The third index indexes into the second element of
5422 the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an array. The
5423 two dimensions of the array are subscripted into, yielding an '<tt>i32</tt>'
5424 type. The '<tt>getelementptr</tt>' instruction returns a pointer to this
5425 element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5426
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5427<p>Note that it is perfectly legal to index partially through a structure,
5428 returning a pointer to an inner element. Because of this, the LLVM code for
5429 the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5430
Bill Wendlinga3495392011-12-13 01:07:07 +0000[diff] [blame]5431<pre class="doc_code">
5432define i32* @foo(%struct.ST* %s) {
5433 %t1 = getelementptr %struct.ST* %s, i32 1 <i>; yields %struct.ST*:%t1</i>
5434 %t2 = getelementptr %struct.ST* %t1, i32 0, i32 2 <i>; yields %struct.RT*:%t2</i>
5435 %t3 = getelementptr %struct.RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
5436 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
5437 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
5438 ret i32* %t5
5439}
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5440</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]5441
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5442<p>If the <tt>inbounds</tt> keyword is present, the result value of the
Dan Gohmanbfb056d2011-12-06 03:18:47 +0000[diff] [blame]5443 <tt>getelementptr</tt> is a <a href="#poisonvalues">poison value</a> if the
Dan Gohman27ef9972010-04-23 15:23:32 +0000[diff] [blame]5444 base pointer is not an <i>in bounds</i> address of an allocated object,
5445 or if any of the addresses that would be formed by successive addition of
5446 the offsets implied by the indices to the base address with infinitely
Eli Friedman266246c2011-08-12 23:37:55 +0000[diff] [blame]5447 precise signed arithmetic are not an <i>in bounds</i> address of that
5448 allocated object. The <i>in bounds</i> addresses for an allocated object
5449 are all the addresses that point into the object, plus the address one
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5450 byte past the end.
5451 In cases where the base is a vector of pointers the <tt>inbounds</tt> keyword
5452 applies to each of the computations element-wise. </p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5453
5454<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]5455 the base address with silently-wrapping two's complement arithmetic. If the
5456 offsets have a different width from the pointer, they are sign-extended or
5457 truncated to the width of the pointer. The result value of the
5458 <tt>getelementptr</tt> may be outside the object pointed to by the base
5459 pointer. The result value may not necessarily be used to access memory
5460 though, even if it happens to point into allocated storage. See the
5461 <a href="#pointeraliasing">Pointer Aliasing Rules</a> section for more
5462 information.</p>
Dan Gohmandd8004d2009-07-27 21:53:46 +0000[diff] [blame]5463
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5464<p>The getelementptr instruction is often confusing. For some more insight into
5465 how it works, see <a href="GetElementPtr.html">the getelementptr FAQ</a>.</p>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]5466
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]5467<h5>Example:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5468<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5469 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijmane49d0bc2008-10-13 13:44:15 +0000[diff] [blame]5470 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
5471 <i>; yields i8*:vptr</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5472 %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]5473 <i>; yields i8*:eptr</i>
5474 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Sanjiv Gupta9f805c22009-04-25 07:27:44 +0000[diff] [blame]5475 <i>; yields i32*:iptr</i>
Sanjiv Gupta16ffa802009-04-24 16:38:13 +0000[diff] [blame]5476 %iptr = getelementptr [10 x i32]* @arr, i16 0, i16 0
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5477</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5478
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5479<p>In cases where the pointer argument is a vector of pointers, only a
5480 single index may be used, and the number of vector elements has to be
5481 the same. For example: </p>
5482<pre class="doc_code">
5483 %A = getelementptr <4 x i8*> %ptrs, <4 x i64> %offsets,
5484</pre>
5485
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]5486</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]5487
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5488</div>
5489
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5490<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5491<h3>
5492 <a name="convertops">Conversion Operations</a>
5493</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5494
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5495<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5496
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5497<p>The instructions in this category are the conversion instructions (casting)
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5498 which all take a single operand and a type. They perform various bit
5499 conversions on the operand.</p>
5500
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5501<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5502<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5503 <a name="i_trunc">'<tt>trunc .. 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>
5510 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5511</pre>
5512
5513<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5514<p>The '<tt>trunc</tt>' instruction truncates its operand to the
5515 type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5516
5517<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5518<p>The '<tt>trunc</tt>' instruction takes a value to trunc, and a type to trunc it to.
5519 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5520 of the same number of integers.
5521 The bit size of the <tt>value</tt> must be larger than
5522 the bit size of the destination type, <tt>ty2</tt>.
5523 Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5524
5525<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5526<p>The '<tt>trunc</tt>' instruction truncates the high order bits
5527 in <tt>value</tt> and converts the remaining bits to <tt>ty2</tt>. Since the
5528 source size must be larger than the destination size, <tt>trunc</tt> cannot
5529 be a <i>no-op cast</i>. It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5530
5531<h5>Example:</h5>
5532<pre>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5533 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
5534 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
5535 %Z = trunc i32 122 to i1 <i>; yields i1:false</i>
5536 %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]5537</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5538
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5539</div>
5540
5541<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5542<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5543 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5544</h4>
5545
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5546<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5547
5548<h5>Syntax:</h5>
5549<pre>
5550 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5551</pre>
5552
5553<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5554<p>The '<tt>zext</tt>' instruction zero extends its operand to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5555 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5556
5557
5558<h5>Arguments:</h5>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5559<p>The '<tt>zext</tt>' instruction takes a value to cast, and a type to cast it to.
5560 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5561 of the same number of integers.
5562 The bit size of the <tt>value</tt> must be smaller than
5563 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5564 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5565
5566<h5>Semantics:</h5>
5567<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]5568 bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5569
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]5570<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]5571
5572<h5>Example:</h5>
5573<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5574 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5575 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Nadav Rotemed9b9342011-02-20 12:37:50 +0000[diff] [blame]5576 %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]5577</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5578
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5579</div>
5580
5581<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5582<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5583 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5584</h4>
5585
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5586<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5587
5588<h5>Syntax:</h5>
5589<pre>
5590 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5591</pre>
5592
5593<h5>Overview:</h5>
5594<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
5595
5596<h5>Arguments:</h5>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5597<p>The '<tt>sext</tt>' instruction takes a value to cast, and a type to cast it to.
5598 Both types must be of <a href="#t_integer">integer</a> types, or vectors
5599 of the same number of integers.
5600 The bit size of the <tt>value</tt> must be smaller than
5601 the bit size of the destination type,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5602 <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5603
5604<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5605<p>The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
5606 bit (highest order bit) of the <tt>value</tt> until it reaches the bit size
5607 of the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5608
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5609<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]5610
5611<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5612<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5613 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5614 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Nadav Rotem8c20ec52011-02-24 21:01:34 +0000[diff] [blame]5615 %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]5616</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5617
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5618</div>
5619
5620<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5621<h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5622 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5623</h4>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5624
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5625<div>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5626
5627<h5>Syntax:</h5>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5628<pre>
5629 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5630</pre>
5631
5632<h5>Overview:</h5>
5633<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5634 <tt>ty2</tt>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5635
5636<h5>Arguments:</h5>
5637<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5638 point</a> value to cast and a <a href="#t_floating">floating point</a> type
5639 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]5640 <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]5641 <i>no-op cast</i>.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5642
5643<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5644<p>The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5645 <a href="#t_floating">floating point</a> type to a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5646 <a href="#t_floating">floating point</a> type. If the value cannot fit
5647 within the destination type, <tt>ty2</tt>, then the results are
5648 undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5649
5650<h5>Example:</h5>
5651<pre>
5652 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
5653 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
5654</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5655
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]5656</div>
5657
5658<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5659<h4>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5660 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5661</h4>
5662
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5663<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5664
5665<h5>Syntax:</h5>
5666<pre>
5667 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5668</pre>
5669
5670<h5>Overview:</h5>
5671<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]5672 floating point value.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5673
5674<h5>Arguments:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5675<p>The '<tt>fpext</tt>' instruction takes a
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5676 <a href="#t_floating">floating point</a> <tt>value</tt> to cast, and
5677 a <a href="#t_floating">floating point</a> type to cast it to. The source
5678 type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5679
5680<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5681<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5682 <a href="#t_floating">floating point</a> type to a larger
5683 <a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
5684 used to make a <i>no-op cast</i> because it always changes bits. Use
5685 <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]5686
5687<h5>Example:</h5>
5688<pre>
Nick Lewycky5bb3ece2011-03-31 18:20:19 +0000[diff] [blame]5689 %X = fpext float 3.125 to double <i>; yields double:3.125000e+00</i>
5690 %Y = fpext double %X to fp128 <i>; yields fp128:0xL00000000000000004000900000000000</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5691</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5692
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5693</div>
5694
5695<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5696<h4>
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]5697 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5698</h4>
5699
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5700<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5701
5702<h5>Syntax:</h5>
5703<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5704 &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]5705</pre>
5706
5707<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5708<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5709 unsigned integer equivalent of type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5710
5711<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5712<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
5713 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5714 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5715 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5716 vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5717
5718<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5719<p>The '<tt>fptoui</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5720 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5721 towards zero) unsigned integer value. If the value cannot fit
5722 in <tt>ty2</tt>, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5723
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5724<h5>Example:</h5>
5725<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]5726 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5727 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5728 %Z = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5729</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5730
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5731</div>
5732
5733<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5734<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5735 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5736</h4>
5737
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5738<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5739
5740<h5>Syntax:</h5>
5741<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5742 &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]5743</pre>
5744
5745<h5>Overview:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5746<p>The '<tt>fptosi</tt>' instruction converts
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5747 <a href="#t_floating">floating point</a> <tt>value</tt> to
5748 type <tt>ty2</tt>.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5749
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5750<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5751<p>The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
5752 scalar or vector <a href="#t_floating">floating point</a> value, and a type
5753 to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
5754 type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
5755 vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5756
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]5757<h5>Semantics:</h5>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]5758<p>The '<tt>fptosi</tt>' instruction converts its
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5759 <a href="#t_floating">floating point</a> operand into the nearest (rounding
5760 towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
5761 the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5762
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]5763<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5764<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]5765 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]5766 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Gabor Greif6a292012009-10-28 09:21:30 +0000[diff] [blame]5767 %Z = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5768</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5769
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5770</div>
5771
5772<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5773<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5774 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5775</h4>
5776
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5777<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5778
5779<h5>Syntax:</h5>
5780<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5781 &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]5782</pre>
5783
5784<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5785<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5786 integer and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5787
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5788<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5789<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]5790 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5791 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5792 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5793 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5794
5795<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5796<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5797 integer quantity and converts it to the corresponding floating point
5798 value. If the value cannot fit in the floating point value, the results are
5799 undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5800
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5801<h5>Example:</h5>
5802<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5803 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5804 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5805</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5806
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5807</div>
5808
5809<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5810<h4>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5811 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5812</h4>
5813
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5814<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5815
5816<h5>Syntax:</h5>
5817<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]5818 &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]5819</pre>
5820
5821<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5822<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed integer
5823 and converts that value to the <tt>ty2</tt> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5824
5825<h5>Arguments:</h5>
Nate Begemanb348d182007-11-17 03:58:34 +0000[diff] [blame]5826<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]5827 scalar or vector <a href="#t_integer">integer</a> value, and a type to cast
5828 it to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
5829 type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
5830 floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5831
5832<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5833<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed integer
5834 quantity and converts it to the corresponding floating point value. If the
5835 value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5836
5837<h5>Example:</h5>
5838<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5839 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]5840 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5841</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5842
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5843</div>
5844
5845<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5846<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5847 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5848</h4>
5849
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5850<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5851
5852<h5>Syntax:</h5>
5853<pre>
5854 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5855</pre>
5856
5857<h5>Overview:</h5>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5858<p>The '<tt>ptrtoint</tt>' instruction converts the pointer or a vector of
5859 pointers <tt>value</tt> to
5860 the integer (or vector of integers) type <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5861
5862<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5863<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5864 must be a a value of type <a href="#t_pointer">pointer</a> or a vector of
5865 pointers, and a type to cast it to
5866 <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> or a vector
5867 of integers type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5868
5869<h5>Semantics:</h5>
5870<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5871 <tt>ty2</tt> by interpreting the pointer value as an integer and either
5872 truncating or zero extending that value to the size of the integer type. If
5873 <tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
5874 <tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
5875 are the same size, then nothing is done (<i>no-op cast</i>) other than a type
5876 change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5877
5878<h5>Example:</h5>
5879<pre>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5880 %X = ptrtoint i32* %P to i8 <i>; yields truncation on 32-bit architecture</i>
5881 %Y = ptrtoint i32* %P to i64 <i>; yields zero extension on 32-bit architecture</i>
5882 %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]5883</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5884
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5885</div>
5886
5887<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5888<h4>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5889 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5890</h4>
5891
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5892<div>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5893
5894<h5>Syntax:</h5>
5895<pre>
5896 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
5897</pre>
5898
5899<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5900<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to a
5901 pointer type, <tt>ty2</tt>.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5902
5903<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]5904<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5905 value to cast, and a type to cast it to, which must be a
5906 <a href="#t_pointer">pointer</a> type.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5907
5908<h5>Semantics:</h5>
5909<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5910 <tt>ty2</tt> by applying either a zero extension or a truncation depending on
5911 the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
5912 size of a pointer then a truncation is done. If <tt>value</tt> is smaller
5913 than the size of a pointer then a zero extension is done. If they are the
5914 same size, nothing is done (<i>no-op cast</i>).</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5915
5916<h5>Example:</h5>
5917<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5918 %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]5919 %Y = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
5920 %Z = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5921 %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]5922</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5923
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]5924</div>
5925
5926<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5927<h4>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5928 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5929</h4>
5930
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5931<div>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5932
5933<h5>Syntax:</h5>
5934<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5935 &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]5936</pre>
5937
5938<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5939<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5940 <tt>ty2</tt> without changing any bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5941
5942<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5943<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be a
5944 non-aggregate first class value, and a type to cast it to, which must also be
5945 a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes
5946 of <tt>value</tt> and the destination type, <tt>ty2</tt>, must be
5947 identical. If the source type is a pointer, the destination type must also be
5948 a pointer. This instruction supports bitwise conversion of vectors to
5949 integers and to vectors of other types (as long as they have the same
5950 size).</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5951
5952<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]5953<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5954 <tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
5955 this conversion. The conversion is done as if the <tt>value</tt> had been
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5956 stored to memory and read back as type <tt>ty2</tt>.
5957 Pointer (or vector of pointers) types may only be converted to other pointer
5958 (or vector of pointers) types with this instruction. To convert
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5959 pointers to other types, use the <a href="#i_inttoptr">inttoptr</a> or
5960 <a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]5961
5962<h5>Example:</h5>
5963<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]5964 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]5965 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5966 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
5967 %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]5968</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5969
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5970</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]5971
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5972</div>
5973
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]5974<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5975<h3>
5976 <a name="otherops">Other Operations</a>
5977</h3>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5978
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5979<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5980
5981<p>The instructions in this category are the "miscellaneous" instructions, which
5982 defy better classification.</p>
5983
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5984<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]5985<h4>
5986 <a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
5987</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5988
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]5989<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5990
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5991<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5992<pre>
5993 &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]5994</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5995
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]5996<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]5997<p>The '<tt>icmp</tt>' instruction returns a boolean value or a vector of
Nadav Rotem16087692011-12-05 06:29:09 +0000[diff] [blame]5998 boolean values based on comparison of its two integer, integer vector,
5999 pointer, or pointer vector operands.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6000
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6001<h5>Arguments:</h5>
6002<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6003 the condition code indicating the kind of comparison to perform. It is not a
6004 value, just a keyword. The possible condition code are:</p>
6005
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6006<ol>
6007 <li><tt>eq</tt>: equal</li>
6008 <li><tt>ne</tt>: not equal </li>
6009 <li><tt>ugt</tt>: unsigned greater than</li>
6010 <li><tt>uge</tt>: unsigned greater or equal</li>
6011 <li><tt>ult</tt>: unsigned less than</li>
6012 <li><tt>ule</tt>: unsigned less or equal</li>
6013 <li><tt>sgt</tt>: signed greater than</li>
6014 <li><tt>sge</tt>: signed greater or equal</li>
6015 <li><tt>slt</tt>: signed less than</li>
6016 <li><tt>sle</tt>: signed less or equal</li>
6017</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6018
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]6019<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6020 <a href="#t_pointer">pointer</a> or integer <a href="#t_vector">vector</a>
6021 typed. They must also be identical types.</p>
6022
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6023<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6024<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to the
6025 condition code given as <tt>cond</tt>. The comparison performed always yields
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]6026 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]6027 result, as follows:</p>
6028
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6029<ol>
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6030 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6031 <tt>false</tt> otherwise. No sign interpretation is necessary or
6032 performed.</li>
6033
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6034 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6035 <tt>false</tt> otherwise. No sign interpretation is necessary or
6036 performed.</li>
6037
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6038 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6039 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
6040
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6041 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6042 <tt>true</tt> if <tt>op1</tt> is greater than or equal
6043 to <tt>op2</tt>.</li>
6044
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6045 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6046 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
6047
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6048 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6049 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
6050
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6051 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6052 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
6053
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6054 <li><tt>sge</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6055 <tt>true</tt> if <tt>op1</tt> is greater than or equal
6056 to <tt>op2</tt>.</li>
6057
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6058 <li><tt>slt</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6059 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
6060
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6061 <li><tt>sle</tt>: interprets the operands as signed values and yields
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6062 <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]6063</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6064
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6065<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6066 values are compared as if they were integers.</p>
6067
6068<p>If the operands are integer vectors, then they are compared element by
6069 element. The result is an <tt>i1</tt> vector with the same number of elements
6070 as the values being compared. Otherwise, the result is an <tt>i1</tt>.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6071
6072<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6073<pre>
6074 &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]6075 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
6076 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
6077 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
6078 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
6079 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6080</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6081
6082<p>Note that the code generator does not yet support vector types with
6083 the <tt>icmp</tt> instruction.</p>
6084
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6085</div>
6086
6087<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6088<h4>
6089 <a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
6090</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6091
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6092<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6093
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6094<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6095<pre>
6096 &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]6097</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6098
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6099<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6100<p>The '<tt>fcmp</tt>' instruction returns a boolean value or vector of boolean
6101 values based on comparison of its operands.</p>
6102
6103<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]6104(<a href="#t_integer"><tt>i1</tt></a>).</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6105
6106<p>If the operands are floating point vectors, then the result type is a vector
6107 of boolean with the same number of elements as the operands being
6108 compared.</p>
6109
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6110<h5>Arguments:</h5>
6111<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6112 the condition code indicating the kind of comparison to perform. It is not a
6113 value, just a keyword. The possible condition code are:</p>
6114
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6115<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6116 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6117 <li><tt>oeq</tt>: ordered and equal</li>
6118 <li><tt>ogt</tt>: ordered and greater than </li>
6119 <li><tt>oge</tt>: ordered and greater than or equal</li>
6120 <li><tt>olt</tt>: ordered and less than </li>
6121 <li><tt>ole</tt>: ordered and less than or equal</li>
6122 <li><tt>one</tt>: ordered and not equal</li>
6123 <li><tt>ord</tt>: ordered (no nans)</li>
6124 <li><tt>ueq</tt>: unordered or equal</li>
6125 <li><tt>ugt</tt>: unordered or greater than </li>
6126 <li><tt>uge</tt>: unordered or greater than or equal</li>
6127 <li><tt>ult</tt>: unordered or less than </li>
6128 <li><tt>ule</tt>: unordered or less than or equal</li>
6129 <li><tt>une</tt>: unordered or not equal</li>
6130 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6131 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6132</ol>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6133
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6134<p><i>Ordered</i> means that neither operand is a QNAN while
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6135 <i>unordered</i> means that either operand may be a QNAN.</p>
6136
6137<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be either
6138 a <a href="#t_floating">floating point</a> type or
6139 a <a href="#t_vector">vector</a> of floating point type. They must have
6140 identical types.</p>
6141
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6142<h5>Semantics:</h5>
Gabor Greiffb224a22008-08-07 21:46:00 +0000[diff] [blame]6143<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6144 according to the condition code given as <tt>cond</tt>. If the operands are
6145 vectors, then the vectors are compared element by element. Each comparison
Nick Lewyckyec38da42009-09-27 00:45:11 +0000[diff] [blame]6146 performed always yields an <a href="#t_integer">i1</a> result, as
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6147 follows:</p>
6148
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6149<ol>
6150 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6151
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6152 <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]6153 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
6154
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6155 <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]6156 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6157
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6158 <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]6159 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
6160
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6161 <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]6162 <tt>op1</tt> is less than <tt>op2</tt>.</li>
6163
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6164 <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]6165 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
6166
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6167 <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]6168 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
6169
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6170 <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]6171
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6172 <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]6173 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
6174
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6175 <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]6176 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
6177
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6178 <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]6179 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
6180
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6181 <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]6182 <tt>op1</tt> is less than <tt>op2</tt>.</li>
6183
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6184 <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]6185 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
6186
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6187 <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]6188 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
6189
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]6190 <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]6191
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6192 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
6193</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6194
6195<h5>Example:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6196<pre>
6197 &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]6198 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
6199 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
6200 &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]6201</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6202
6203<p>Note that the code generator does not yet support vector types with
6204 the <tt>fcmp</tt> instruction.</p>
6205
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]6206</div>
6207
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6208<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6209<h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6210 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6211</h4>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6212
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6213<div>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6214
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6215<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6216<pre>
6217 &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...
6218</pre>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6219
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6220<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6221<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in the
6222 SSA graph representing the function.</p>
6223
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6224<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6225<p>The type of the incoming values is specified with the first type field. After
6226 this, the '<tt>phi</tt>' instruction takes a list of pairs as arguments, with
6227 one pair for each predecessor basic block of the current block. Only values
6228 of <a href="#t_firstclass">first class</a> type may be used as the value
6229 arguments to the PHI node. Only labels may be used as the label
6230 arguments.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6231
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6232<p>There must be no non-phi instructions between the start of a basic block and
6233 the PHI instructions: i.e. PHI instructions must be first in a basic
6234 block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6235
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6236<p>For the purposes of the SSA form, the use of each incoming value is deemed to
6237 occur on the edge from the corresponding predecessor block to the current
6238 block (but after any definition of an '<tt>invoke</tt>' instruction's return
6239 value on the same edge).</p>
Jay Foadd2449092009-06-03 10:20:10 +0000[diff] [blame]6240
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6241<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6242<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6243 specified by the pair corresponding to the predecessor basic block that
6244 executed just prior to the current block.</p>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6245
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6246<h5>Example:</h5>
Chris Lattner5568e942008-05-20 20:48:21 +0000[diff] [blame]6247<pre>
6248Loop: ; Infinite loop that counts from 0 on up...
6249 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
6250 %nextindvar = add i32 %indvar, 1
6251 br label %Loop
6252</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6253
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]6254</div>
6255
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6256<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6257<h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6258 <a name="i_select">'<tt>select</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6259</h4>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6260
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6261<div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6262
6263<h5>Syntax:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6264<pre>
Dan Gohmanf72fb672008-09-09 01:02:47 +0000[diff] [blame]6265 &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>
6266
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6267 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6268</pre>
6269
6270<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6271<p>The '<tt>select</tt>' instruction is used to choose one value based on a
6272 condition, without branching.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6273
6274
6275<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6276<p>The '<tt>select</tt>' instruction requires an 'i1' value or a vector of 'i1'
6277 values indicating the condition, and two values of the
6278 same <a href="#t_firstclass">first class</a> type. If the val1/val2 are
6279 vectors and the condition is a scalar, then entire vectors are selected, not
6280 individual elements.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6281
6282<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6283<p>If the condition is an i1 and it evaluates to 1, the instruction returns the
6284 first value argument; otherwise, it returns the second value argument.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6285
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6286<p>If the condition is a vector of i1, then the value arguments must be vectors
6287 of the same size, and the selection is done element by element.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6288
6289<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6290<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]6291 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6292</pre>
Dan Gohmanc766f722009-01-22 01:39:38 +0000[diff] [blame]6293
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]6294</div>
6295
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]6296<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6297<h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6298 <a name="i_call">'<tt>call</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6299</h4>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6300
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6301<div>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6302
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6303<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6304<pre>
Devang Patel307e8ab2008-10-07 17:48:33 +0000[diff] [blame]6305 &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]6306</pre>
6307
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6308<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6309<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6310
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6311<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6312<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6313
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6314<ol>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6315 <li>The optional "tail" marker indicates that the callee function does not
6316 access any allocas or varargs in the caller. Note that calls may be
6317 marked "tail" even if they do not occur before
6318 a <a href="#i_ret"><tt>ret</tt></a> instruction. If the "tail" marker is
6319 present, the function call is eligible for tail call optimization,
6320 but <a href="CodeGenerator.html#tailcallopt">might not in fact be
Evan Chengdc444e92010-03-08 21:05:02 +0000[diff] [blame]6321 optimized into a jump</a>. The code generator may optimize calls marked
6322 "tail" with either 1) automatic <a href="CodeGenerator.html#sibcallopt">
6323 sibling call optimization</a> when the caller and callee have
6324 matching signatures, or 2) forced tail call optimization when the
6325 following extra requirements are met:
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6326 <ul>
6327 <li>Caller and callee both have the calling
6328 convention <tt>fastcc</tt>.</li>
6329 <li>The call is in tail position (ret immediately follows call and ret
6330 uses value of call or is void).</li>
6331 <li>Option <tt>-tailcallopt</tt> is enabled,
Dan Gohmanfbbee8d2010-03-02 01:08:11 +0000[diff] [blame]6332 or <code>llvm::GuaranteedTailCallOpt</code> is <code>true</code>.</li>
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6333 <li><a href="CodeGenerator.html#tailcallopt">Platform specific
6334 constraints are met.</a></li>
6335 </ul>
6336 </li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6337
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6338 <li>The optional "cconv" marker indicates which <a href="#callingconv">calling
6339 convention</a> the call should use. If none is specified, the call
Jeffrey Yasskin95fa80a2010-01-09 19:44:16 +0000[diff] [blame]6340 defaults to using C calling conventions. The calling convention of the
6341 call must match the calling convention of the target function, or else the
6342 behavior is undefined.</li>
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6343
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6344 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
6345 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>', and
6346 '<tt>inreg</tt>' attributes are valid here.</li>
6347
6348 <li>'<tt>ty</tt>': the type of the call instruction itself which is also the
6349 type of the return value. Functions that return no value are marked
6350 <tt><a href="#t_void">void</a></tt>.</li>
6351
6352 <li>'<tt>fnty</tt>': shall be the signature of the pointer to function value
6353 being invoked. The argument types must match the types implied by this
6354 signature. This type can be omitted if the function is not varargs and if
6355 the function type does not return a pointer to a function.</li>
6356
6357 <li>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
6358 be invoked. In most cases, this is a direct function invocation, but
6359 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
6360 to function value.</li>
6361
6362 <li>'<tt>function args</tt>': argument list whose types match the function
Chris Lattner0724fbd2010-03-02 06:36:51 +0000[diff] [blame]6363 signature argument types and parameter attributes. All arguments must be
6364 of <a href="#t_firstclass">first class</a> type. If the function
6365 signature indicates the function accepts a variable number of arguments,
6366 the extra arguments can be specified.</li>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6367
6368 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
6369 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
6370 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]6371</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6372
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6373<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6374<p>The '<tt>call</tt>' instruction is used to cause control flow to transfer to
6375 a specified function, with its incoming arguments bound to the specified
6376 values. Upon a '<tt><a href="#i_ret">ret</a></tt>' instruction in the called
6377 function, control flow continues with the instruction after the function
6378 call, and the return value of the function is bound to the result
6379 argument.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6380
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]6381<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6382<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]6383 %retval = call i32 @test(i32 %argc)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6384 call i32 (i8*, ...)* @printf(i8* %msg, i32 12, i8 42) <i>; yields i32</i>
Chris Lattner772fccf2008-03-21 17:24:17 +0000[diff] [blame]6385 %X = tail call i32 @foo() <i>; yields i32</i>
6386 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
6387 call void %foo(i8 97 signext)
Devang Patelc3fc6df2008-03-10 20:49:15 +0000[diff] [blame]6388
6389 %struct.A = type { i32, i8 }
Devang Patelf642f472008-10-06 18:50:38 +0000[diff] [blame]6390 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmanb1e6b962008-10-04 19:00:07 +0000[diff] [blame]6391 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
6392 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner85a350f2008-10-08 06:26:11 +0000[diff] [blame]6393 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmancb73d192008-10-07 10:03:45 +0000[diff] [blame]6394 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]6395</pre>
6396
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6397<p>llvm treats calls to some functions with names and arguments that match the
Dale Johannesen9f8380b2009-09-25 17:04:42 +0000[diff] [blame]6398standard C99 library as being the C99 library functions, and may perform
6399optimizations or generate code for them under that assumption. This is
6400something we'd like to change in the future to provide better support for
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]6401freestanding environments and non-C-based languages.</p>
Dale Johannesen07de8d12009-09-24 18:38:21 +0000[diff] [blame]6402
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6403</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6404
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6405<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6406<h4>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]6407 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6408</h4>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6409
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6410<div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6411
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6412<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6413<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6414 &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]6415</pre>
6416
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6417<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6418<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6419 the "variable argument" area of a function call. It is used to implement the
6420 <tt>va_arg</tt> macro in C.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6421
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6422<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6423<p>This instruction takes a <tt>va_list*</tt> value and the type of the
6424 argument. It returns a value of the specified argument type and increments
6425 the <tt>va_list</tt> to point to the next argument. The actual type
6426 of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6427
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6428<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6429<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified type
6430 from the specified <tt>va_list</tt> and causes the <tt>va_list</tt> to point
6431 to the next argument. For more information, see the variable argument
6432 handling <a href="#int_varargs">Intrinsic Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6433
6434<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]6435 take a variable number of arguments, for example, the <tt>vfprintf</tt>
6436 function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6437
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6438<p><tt>va_arg</tt> is an LLVM instruction instead of
6439 an <a href="#intrinsics">intrinsic function</a> because it takes a type as an
6440 argument.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6441
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]6442<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]6443<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
6444
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6445<p>Note that the code generator does not yet fully support va_arg on many
6446 targets. Also, it does not currently support va_arg with aggregate types on
6447 any target.</p>
Dan Gohmanf3e60bd2009-01-12 23:12:39 +0000[diff] [blame]6448
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6449</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6450
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6451<!-- _______________________________________________________________________ -->
6452<h4>
6453 <a name="i_landingpad">'<tt>landingpad</tt>' Instruction</a>
6454</h4>
6455
6456<div>
6457
6458<h5>Syntax:</h5>
6459<pre>
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6460 &lt;resultval&gt; = landingpad &lt;resultty&gt; personality &lt;type&gt; &lt;pers_fn&gt; &lt;clause&gt;+
6461 &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]6462
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6463 &lt;clause&gt; := catch &lt;type&gt; &lt;value&gt;
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6464 &lt;clause&gt; := filter &lt;array constant type&gt; &lt;array constant&gt;
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6465</pre>
6466
6467<h5>Overview:</h5>
6468<p>The '<tt>landingpad</tt>' instruction is used by
6469 <a href="ExceptionHandling.html#overview">LLVM's exception handling
6470 system</a> to specify that a basic block is a landing pad &mdash; one where
6471 the exception lands, and corresponds to the code found in the
6472 <i><tt>catch</tt></i> portion of a <i><tt>try/catch</tt></i> sequence. It
6473 defines values supplied by the personality function (<tt>pers_fn</tt>) upon
6474 re-entry to the function. The <tt>resultval</tt> has the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6475 type <tt>resultty</tt>.</p>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6476
6477<h5>Arguments:</h5>
6478<p>This instruction takes a <tt>pers_fn</tt> value. This is the personality
6479 function associated with the unwinding mechanism. The optional
6480 <tt>cleanup</tt> flag indicates that the landing pad block is a cleanup.</p>
6481
6482<p>A <tt>clause</tt> begins with the clause type &mdash; <tt>catch</tt>
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6483 or <tt>filter</tt> &mdash; and contains the global variable representing the
6484 "type" that may be caught or filtered respectively. Unlike the
6485 <tt>catch</tt> clause, the <tt>filter</tt> clause takes an array constant as
6486 its argument. Use "<tt>[0 x i8**] undef</tt>" for a filter which cannot
6487 throw. The '<tt>landingpad</tt>' instruction must contain <em>at least</em>
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6488 one <tt>clause</tt> or the <tt>cleanup</tt> flag.</p>
6489
6490<h5>Semantics:</h5>
6491<p>The '<tt>landingpad</tt>' instruction defines the values which are set by the
6492 personality function (<tt>pers_fn</tt>) upon re-entry to the function, and
6493 therefore the "result type" of the <tt>landingpad</tt> instruction. As with
6494 calling conventions, how the personality function results are represented in
6495 LLVM IR is target specific.</p>
6496
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6497<p>The clauses are applied in order from top to bottom. If two
6498 <tt>landingpad</tt> instructions are merged together through inlining, the
Duncan Sands8d6796b2012-01-13 19:59:16 +0000[diff] [blame]6499 clauses from the calling function are appended to the list of clauses.
6500 When the call stack is being unwound due to an exception being thrown, the
6501 exception is compared against each <tt>clause</tt> in turn. If it doesn't
6502 match any of the clauses, and the <tt>cleanup</tt> flag is not set, then
6503 unwinding continues further up the call stack.</p>
Bill Wendlingb7a01352011-08-03 17:17:06 +0000[diff] [blame]6504
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6505<p>The <tt>landingpad</tt> instruction has several restrictions:</p>
6506
6507<ul>
6508 <li>A landing pad block is a basic block which is the unwind destination of an
6509 '<tt>invoke</tt>' instruction.</li>
6510 <li>A landing pad block must have a '<tt>landingpad</tt>' instruction as its
6511 first non-PHI instruction.</li>
6512 <li>There can be only one '<tt>landingpad</tt>' instruction within the landing
6513 pad block.</li>
6514 <li>A basic block that is not a landing pad block may not include a
6515 '<tt>landingpad</tt>' instruction.</li>
6516 <li>All '<tt>landingpad</tt>' instructions in a function must have the same
6517 personality function.</li>
6518</ul>
6519
6520<h5>Example:</h5>
6521<pre>
6522 ;; A landing pad which can catch an integer.
6523 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6524 catch i8** @_ZTIi
6525 ;; A landing pad that is a cleanup.
6526 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6527 cleanup
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6528 ;; A landing pad which can catch an integer and can only throw a double.
6529 %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
6530 catch i8** @_ZTIi
Bill Wendlinge6e88262011-08-12 20:24:12 +0000[diff] [blame]6531 filter [1 x i8**] [@_ZTId]
Bill Wendlingf78faf82011-08-02 21:52:38 +0000[diff] [blame]6532</pre>
6533
6534</div>
6535
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6536</div>
6537
6538</div>
6539
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6540<!-- *********************************************************************** -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6541<h2><a name="intrinsics">Intrinsic Functions</a></h2>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]6542<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6543
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6544<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6545
6546<p>LLVM supports the notion of an "intrinsic function". These functions have
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6547 well known names and semantics and are required to follow certain
6548 restrictions. Overall, these intrinsics represent an extension mechanism for
6549 the LLVM language that does not require changing all of the transformations
6550 in LLVM when adding to the language (or the bitcode reader/writer, the
6551 parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6552
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]6553<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6554 prefix is reserved in LLVM for intrinsic names; thus, function names may not
6555 begin with this prefix. Intrinsic functions must always be external
6556 functions: you cannot define the body of intrinsic functions. Intrinsic
6557 functions may only be used in call or invoke instructions: it is illegal to
6558 take the address of an intrinsic function. Additionally, because intrinsic
6559 functions are part of the LLVM language, it is required if any are added that
6560 they be documented here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6561
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6562<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents a
6563 family of functions that perform the same operation but on different data
6564 types. Because LLVM can represent over 8 million different integer types,
6565 overloading is used commonly to allow an intrinsic function to operate on any
6566 integer type. One or more of the argument types or the result type can be
6567 overloaded to accept any integer type. Argument types may also be defined as
6568 exactly matching a previous argument's type or the result type. This allows
6569 an intrinsic function which accepts multiple arguments, but needs all of them
6570 to be of the same type, to only be overloaded with respect to a single
6571 argument or the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6572
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6573<p>Overloaded intrinsics will have the names of its overloaded argument types
6574 encoded into its function name, each preceded by a period. Only those types
6575 which are overloaded result in a name suffix. Arguments whose type is matched
6576 against another type do not. For example, the <tt>llvm.ctpop</tt> function
6577 can take an integer of any width and returns an integer of exactly the same
6578 integer width. This leads to a family of functions such as
6579 <tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29
6580 %val)</tt>. Only one type, the return type, is overloaded, and only one type
6581 suffix is required. Because the argument's type is matched against the return
6582 type, it does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]6583
Eric Christopher6c7e8a02009-12-05 02:46:03 +0000[diff] [blame]6584<p>To learn how to add an intrinsic function, please see the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6585 <a href="ExtendingLLVM.html">Extending LLVM Guide</a>.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6586
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6587<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6588<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6589 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6590</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6591
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6592<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6593
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6594<p>Variable argument support is defined in LLVM with
6595 the <a href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
6596 intrinsic functions. These functions are related to the similarly named
6597 macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6598
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6599<p>All of these functions operate on arguments that use a target-specific value
6600 type "<tt>va_list</tt>". The LLVM assembly language reference manual does
6601 not define what this type is, so all transformations should be prepared to
6602 handle these functions regardless of the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6603
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]6604<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]6605 instruction and the variable argument handling intrinsic functions are
6606 used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6607
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]6608<pre class="doc_code">
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6609define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6610 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6611 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6612 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6613 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6614
6615 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6616 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6617
6618 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6619 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6620 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6621 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6622 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6623
6624 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6625 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]6626 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6627}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]6628
6629declare void @llvm.va_start(i8*)
6630declare void @llvm.va_copy(i8*, i8*)
6631declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6632</pre>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6633
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6634<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6635<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6636 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6637</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6638
6639
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6640<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6641
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6642<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6643<pre>
6644 declare void %llvm.va_start(i8* &lt;arglist&gt;)
6645</pre>
6646
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6647<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6648<p>The '<tt>llvm.va_start</tt>' intrinsic initializes <tt>*&lt;arglist&gt;</tt>
6649 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]6650
6651<h5>Arguments:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6652<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]6653
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6654<h5>Semantics:</h5>
Dan Gohman0e451ce2008-10-14 16:51:45 +0000[diff] [blame]6655<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]6656 macro available in C. In a target-dependent way, it initializes
6657 the <tt>va_list</tt> element to which the argument points, so that the next
6658 call to <tt>va_arg</tt> will produce the first variable argument passed to
6659 the function. Unlike the C <tt>va_start</tt> macro, this intrinsic does not
6660 need to know the last argument of the function as the compiler can figure
6661 that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6662
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6663</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6664
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6665<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6666<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6667 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6668</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6669
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6670<div>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6671
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6672<h5>Syntax:</h5>
6673<pre>
6674 declare void @llvm.va_end(i8* &lt;arglist&gt;)
6675</pre>
6676
6677<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6678<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6679 which has been initialized previously
6680 with <tt><a href="#int_va_start">llvm.va_start</a></tt>
6681 or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6682
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6683<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6684<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]6685
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6686<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6687<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]6688 macro available in C. In a target-dependent way, it destroys
6689 the <tt>va_list</tt> element to which the argument points. Calls
6690 to <a href="#int_va_start"><tt>llvm.va_start</tt></a>
6691 and <a href="#int_va_copy"> <tt>llvm.va_copy</tt></a> must be matched exactly
6692 with calls to <tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]6693
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6694</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6695
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6696<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6697<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6698 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6699</h4>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6700
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6701<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6702
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6703<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6704<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6705 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6706</pre>
6707
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6708<h5>Overview:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6709<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6710 from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6711
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6712<h5>Arguments:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]6713<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]6714 The second argument is a pointer to a <tt>va_list</tt> element to copy
6715 from.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6716
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]6717<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]6718<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]6719 macro available in C. In a target-dependent way, it copies the
6720 source <tt>va_list</tt> element into the destination <tt>va_list</tt>
6721 element. This intrinsic is necessary because
6722 the <tt><a href="#int_va_start"> llvm.va_start</a></tt> intrinsic may be
6723 arbitrarily complex and require, for example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6724
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]6725</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]6726
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6727</div>
6728
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]6729<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6730<h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6731 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6732</h3>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6733
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6734<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6735
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6736<p>LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattnerd3eda892008-08-05 18:29:16 +0000[diff] [blame]6737Collection</a> (GC) requires the implementation and generation of these
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6738intrinsics. These intrinsics allow identification of <a href="#int_gcroot">GC
6739roots on the stack</a>, as well as garbage collector implementations that
6740require <a href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a>
6741barriers. Front-ends for type-safe garbage collected languages should generate
6742these intrinsics to make use of the LLVM garbage collectors. For more details,
6743see <a href="GarbageCollection.html">Accurate Garbage Collection with
6744LLVM</a>.</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6745
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6746<p>The garbage collection intrinsics only operate on objects in the generic
6747 address space (address space zero).</p>
Christopher Lamb303dae92007-12-17 01:00:21 +0000[diff] [blame]6748
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6749<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6750<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6751 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6752</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6753
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6754<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6755
6756<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6757<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6758 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6759</pre>
6760
6761<h5>Overview:</h5>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]6762<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]6763 the code generator, and allows some metadata to be associated with it.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6764
6765<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6766<p>The first argument specifies the address of a stack object that contains the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6767 root pointer. The second pointer (which must be either a constant or a
6768 global value address) contains the meta-data to be associated with the
6769 root.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6770
6771<h5>Semantics:</h5>
Chris Lattner05d67092008-04-24 05:59:56 +0000[diff] [blame]6772<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]6773 location. At compile-time, the code generator generates information to allow
6774 the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
6775 intrinsic may only be used in a function which <a href="#gc">specifies a GC
6776 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6777
6778</div>
6779
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6780<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6781<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6782 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6783</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6784
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6785<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6786
6787<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6788<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6789 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6790</pre>
6791
6792<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6793<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6794 locations, allowing garbage collector implementations that require read
6795 barriers.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6796
6797<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6798<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]6799 allocated from the garbage collector. The first object is a pointer to the
6800 start of the referenced object, if needed by the language runtime (otherwise
6801 null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6802
6803<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6804<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6805 instruction, but may be replaced with substantially more complex code by the
6806 garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
6807 may only be used in a function which <a href="#gc">specifies a GC
6808 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6809
6810</div>
6811
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6812<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6813<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6814 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6815</h4>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6816
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6817<div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6818
6819<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6820<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]6821 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6822</pre>
6823
6824<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6825<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6826 locations, allowing garbage collector implementations that require write
6827 barriers (such as generational or reference counting collectors).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6828
6829<h5>Arguments:</h5>
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]6830<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]6831 object to store it to, and the third is the address of the field of Obj to
6832 store to. If the runtime does not require a pointer to the object, Obj may
6833 be null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6834
6835<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6836<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6837 instruction, but may be replaced with substantially more complex code by the
6838 garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
6839 may only be used in a function which <a href="#gc">specifies a GC
6840 algorithm</a>.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6841
6842</div>
6843
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6844</div>
6845
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]6846<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6847<h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6848 <a name="int_codegen">Code Generator Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6849</h3>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6850
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6851<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6852
6853<p>These intrinsics are provided by LLVM to expose special features that may
6854 only be implemented with code generator support.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6855
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6856<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6857<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6858 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6859</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6860
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6861<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6862
6863<h5>Syntax:</h5>
6864<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]6865 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6866</pre>
6867
6868<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6869<p>The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
6870 target-specific value indicating the return address of the current function
6871 or one of its callers.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6872
6873<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6874<p>The argument to this intrinsic indicates which function to return the address
6875 for. Zero indicates the calling function, one indicates its caller, etc.
6876 The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6877
6878<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6879<p>The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer
6880 indicating the return address of the specified call frame, or zero if it
6881 cannot be identified. The value returned by this intrinsic is likely to be
6882 incorrect or 0 for arguments other than zero, so it should only be used for
6883 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6884
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6885<p>Note that calling this intrinsic does not prevent function inlining or other
6886 aggressive transformations, so the value returned may not be that of the
6887 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6888
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6889</div>
6890
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6891<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6892<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6893 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6894</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6895
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6896<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6897
6898<h5>Syntax:</h5>
6899<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6900 declare i8* @llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6901</pre>
6902
6903<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6904<p>The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
6905 target-specific frame pointer value for the specified stack frame.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6906
6907<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6908<p>The argument to this intrinsic indicates which function to return the frame
6909 pointer for. Zero indicates the calling function, one indicates its caller,
6910 etc. The argument is <b>required</b> to be a constant integer value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6911
6912<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6913<p>The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer
6914 indicating the frame address of the specified call frame, or zero if it
6915 cannot be identified. The value returned by this intrinsic is likely to be
6916 incorrect or 0 for arguments other than zero, so it should only be used for
6917 debugging purposes.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6918
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6919<p>Note that calling this intrinsic does not prevent function inlining or other
6920 aggressive transformations, so the value returned may not be that of the
6921 obvious source-language caller.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6922
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]6923</div>
6924
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6925<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6926<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6927 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6928</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6929
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6930<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6931
6932<h5>Syntax:</h5>
6933<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6934 declare i8* @llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6935</pre>
6936
6937<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6938<p>The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state
6939 of the function stack, for use
6940 with <a href="#int_stackrestore"> <tt>llvm.stackrestore</tt></a>. This is
6941 useful for implementing language features like scoped automatic variable
6942 sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6943
6944<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6945<p>This intrinsic returns a opaque pointer value that can be passed
6946 to <a href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When
6947 an <tt>llvm.stackrestore</tt> intrinsic is executed with a value saved
6948 from <tt>llvm.stacksave</tt>, it effectively restores the state of the stack
6949 to the state it was in when the <tt>llvm.stacksave</tt> intrinsic executed.
6950 In practice, this pops any <a href="#i_alloca">alloca</a> blocks from the
6951 stack that were allocated after the <tt>llvm.stacksave</tt> was executed.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6952
6953</div>
6954
6955<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6956<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6957 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6958</h4>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6959
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6960<div>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6961
6962<h5>Syntax:</h5>
6963<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]6964 declare void @llvm.stackrestore(i8* %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6965</pre>
6966
6967<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6968<p>The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
6969 the function stack to the state it was in when the
6970 corresponding <a href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic
6971 executed. This is useful for implementing language features like scoped
6972 automatic variable sized arrays in C99.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6973
6974<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6975<p>See the description
6976 for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.</p>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6977
6978</div>
6979
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]6980<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6981<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]6982 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]6983</h4>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6984
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]6985<div>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6986
6987<h5>Syntax:</h5>
6988<pre>
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]6989 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]6990</pre>
6991
6992<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6993<p>The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to
6994 insert a prefetch instruction if supported; otherwise, it is a noop.
6995 Prefetches have no effect on the behavior of the program but can change its
6996 performance characteristics.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]6997
6998<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]6999<p><tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the
7000 specifier determining if the fetch should be for a read (0) or write (1),
7001 and <tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Bruno Cardoso Lopes9a767332011-06-14 04:58:37 +0000[diff] [blame]7002 locality, to (3) - extremely local keep in cache. The <tt>cache type</tt>
7003 specifies whether the prefetch is performed on the data (1) or instruction (0)
7004 cache. The <tt>rw</tt>, <tt>locality</tt> and <tt>cache type</tt> arguments
7005 must be constant integers.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]7006
7007<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7008<p>This intrinsic does not modify the behavior of the program. In particular,
7009 prefetches cannot trap and do not produce a value. On targets that support
7010 this intrinsic, the prefetch can provide hints to the processor cache for
7011 better performance.</p>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]7012
7013</div>
7014
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]7015<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7016<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7017 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7018</h4>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]7019
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7020<div>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]7021
7022<h5>Syntax:</h5>
7023<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]7024 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]7025</pre>
7026
7027<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7028<p>The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program
7029 Counter (PC) in a region of code to simulators and other tools. The method
7030 is target specific, but it is expected that the marker will use exported
7031 symbols to transmit the PC of the marker. The marker makes no guarantees
7032 that it will remain with any specific instruction after optimizations. It is
7033 possible that the presence of a marker will inhibit optimizations. The
7034 intended use is to be inserted after optimizations to allow correlations of
7035 simulation runs.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]7036
7037<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7038<p><tt>id</tt> is a numerical id identifying the marker.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]7039
7040<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7041<p>This intrinsic does not modify the behavior of the program. Backends that do
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7042 not support this intrinsic may ignore it.</p>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]7043
7044</div>
7045
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7046<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7047<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7048 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7049</h4>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7050
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7051<div>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7052
7053<h5>Syntax:</h5>
7054<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]7055 declare i64 @llvm.readcyclecounter()
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7056</pre>
7057
7058<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7059<p>The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
7060 counter register (or similar low latency, high accuracy clocks) on those
7061 targets that support it. On X86, it should map to RDTSC. On Alpha, it
7062 should map to RPCC. As the backing counters overflow quickly (on the order
7063 of 9 seconds on alpha), this should only be used for small timings.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7064
7065<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7066<p>When directly supported, reading the cycle counter should not modify any
7067 memory. Implementations are allowed to either return a application specific
7068 value or a system wide value. On backends without support, this is lowered
7069 to a constant 0.</p>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]7070
7071</div>
7072
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7073</div>
7074
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7075<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7076<h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7077 <a name="int_libc">Standard C Library Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7078</h3>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7079
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7080<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7081
7082<p>LLVM provides intrinsics for a few important standard C library functions.
7083 These intrinsics allow source-language front-ends to pass information about
7084 the alignment of the pointer arguments to the code generator, providing
7085 opportunity for more efficient code generation.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7086
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7087<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7088<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7089 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7090</h4>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7091
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7092<div>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7093
7094<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7095<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]7096 integer bit width and for different address spaces. Not all targets support
7097 all bit widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7098
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7099<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7100 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]7101 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7102 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]7103 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7104</pre>
7105
7106<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7107<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
7108 source location to the destination location.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7109
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7110<p>Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7111 intrinsics do not return a value, takes extra alignment/isvolatile arguments
7112 and the pointers can be in specified address spaces.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7113
7114<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7115
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7116<p>The first argument is a pointer to the destination, the second is a pointer
7117 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7118 number of bytes to copy, the fourth argument is the alignment of the
7119 source and destination locations, and the fifth is a boolean indicating a
7120 volatile access.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7121
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7122<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]7123 then the caller guarantees that both the source and destination pointers are
7124 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]7125
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7126<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7127 <tt>llvm.memcpy</tt> call is a <a href="#volatile">volatile operation</a>.
7128 The detailed access behavior is not very cleanly specified and it is unwise
7129 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7130
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7131<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7132
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7133<p>The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the
7134 source location to the destination location, which are not allowed to
7135 overlap. It copies "len" bytes of memory over. If the argument is known to
7136 be aligned to some boundary, this can be specified as the fourth argument,
7137 otherwise it should be set to 0 or 1.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7138
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]7139</div>
7140
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7141<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7142<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7143 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7144</h4>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7145
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7146<div>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7147
7148<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]7149<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]7150 width and for different address space. Not all targets support all bit
7151 widths however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7152
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7153<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7154 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]7155 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7156 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]7157 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7158</pre>
7159
7160<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7161<p>The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the
7162 source location to the destination location. It is similar to the
7163 '<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to
7164 overlap.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7165
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7166<p>Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7167 intrinsics do not return a value, takes extra alignment/isvolatile arguments
7168 and the pointers can be in specified address spaces.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7169
7170<h5>Arguments:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7171
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7172<p>The first argument is a pointer to the destination, the second is a pointer
7173 to the source. The third argument is an integer argument specifying the
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7174 number of bytes to copy, the fourth argument is the alignment of the
7175 source and destination locations, and the fifth is a boolean indicating a
7176 volatile access.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7177
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7178<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]7179 then the caller guarantees that the source and destination pointers are
7180 aligned to that boundary.</p>
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]7181
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7182<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7183 <tt>llvm.memmove</tt> call is a <a href="#volatile">volatile operation</a>.
7184 The detailed access behavior is not very cleanly specified and it is unwise
7185 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7186
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7187<h5>Semantics:</h5>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7188
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7189<p>The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the
7190 source location to the destination location, which may overlap. It copies
7191 "len" bytes of memory over. If the argument is known to be aligned to some
7192 boundary, this can be specified as the fourth argument, otherwise it should
7193 be set to 0 or 1.</p>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7194
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]7195</div>
7196
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7197<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7198<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7199 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7200</h4>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7201
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7202<div>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7203
7204<h5>Syntax:</h5>
Chris Lattner824b9582008-11-21 16:42:48 +0000[diff] [blame]7205<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]7206 width and for different address spaces. However, not all targets support all
7207 bit widths.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7208
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7209<pre>
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7210 declare void @llvm.memset.p0i8.i32(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]7211 i32 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Dan Gohmanfe47aae2010-05-28 17:13:49 +0000[diff] [blame]7212 declare void @llvm.memset.p0i8.i64(i8* &lt;dest&gt;, i8 &lt;val&gt;,
Chris Lattnerff35c3f2010-04-08 00:54:34 +0000[diff] [blame]7213 i64 &lt;len&gt;, i32 &lt;align&gt;, i1 &lt;isvolatile&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7214</pre>
7215
7216<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7217<p>The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a
7218 particular byte value.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7219
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7220<p>Note that, unlike the standard libc function, the <tt>llvm.memset</tt>
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]7221 intrinsic does not return a value and takes extra alignment/volatile
7222 arguments. Also, the destination can be in an arbitrary address space.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7223
7224<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7225<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]7226 byte value with which to fill it, the third argument is an integer argument
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7227 specifying the number of bytes to fill, and the fourth argument is the known
John Criswellcdcbbfc2010-07-30 16:30:28 +0000[diff] [blame]7228 alignment of the destination location.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7229
Dan Gohmanb55a1ee2010-03-01 17:41:39 +0000[diff] [blame]7230<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]7231 then the caller guarantees that the destination pointer is aligned to that
7232 boundary.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7233
Jeffrey Yasskin93e066d2010-04-26 21:21:24 +0000[diff] [blame]7234<p>If the <tt>isvolatile</tt> parameter is <tt>true</tt>, the
7235 <tt>llvm.memset</tt> call is a <a href="#volatile">volatile operation</a>.
7236 The detailed access behavior is not very cleanly specified and it is unwise
7237 to depend on it.</p>
Chris Lattner9f636de2010-04-08 00:53:57 +0000[diff] [blame]7238
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7239<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7240<p>The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting
7241 at the destination location. If the argument is known to be aligned to some
7242 boundary, this can be specified as the fourth argument, otherwise it should
7243 be set to 0 or 1.</p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7244
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]7245</div>
7246
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7247<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7248<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7249 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7250</h4>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7251
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7252<div>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7253
7254<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7255<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
7256 floating point or vector of floating point type. Not all targets support all
7257 types however.</p>
7258
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7259<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]7260 declare float @llvm.sqrt.f32(float %Val)
7261 declare double @llvm.sqrt.f64(double %Val)
7262 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
7263 declare fp128 @llvm.sqrt.f128(fp128 %Val)
7264 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7265</pre>
7266
7267<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7268<p>The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
7269 returning the same value as the libm '<tt>sqrt</tt>' functions would.
7270 Unlike <tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined
7271 behavior for negative numbers other than -0.0 (which allows for better
7272 optimization, because there is no need to worry about errno being
7273 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]7274
7275<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7276<p>The argument and return value are floating point numbers of the same
7277 type.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7278
7279<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7280<p>This function returns the sqrt of the specified operand if it is a
7281 nonnegative floating point number.</p>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7282
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]7283</div>
7284
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7285<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7286<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7287 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7288</h4>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7289
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7290<div>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7291
7292<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7293<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
7294 floating point or vector of floating point type. Not all targets support all
7295 types however.</p>
7296
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7297<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]7298 declare float @llvm.powi.f32(float %Val, i32 %power)
7299 declare double @llvm.powi.f64(double %Val, i32 %power)
7300 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
7301 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
7302 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7303</pre>
7304
7305<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7306<p>The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
7307 specified (positive or negative) power. The order of evaluation of
7308 multiplications is not defined. When a vector of floating point type is
7309 used, the second argument remains a scalar integer value.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7310
7311<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7312<p>The second argument is an integer power, and the first is a value to raise to
7313 that power.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7314
7315<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7316<p>This function returns the first value raised to the second power with an
7317 unspecified sequence of rounding operations.</p>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7318
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]7319</div>
7320
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7321<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7322<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7323 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7324</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7325
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7326<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7327
7328<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7329<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
7330 floating point or vector of floating point type. Not all targets support all
7331 types however.</p>
7332
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7333<pre>
7334 declare float @llvm.sin.f32(float %Val)
7335 declare double @llvm.sin.f64(double %Val)
7336 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
7337 declare fp128 @llvm.sin.f128(fp128 %Val)
7338 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
7339</pre>
7340
7341<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7342<p>The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7343
7344<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7345<p>The argument and return value are floating point numbers of the same
7346 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7347
7348<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7349<p>This function returns the sine of the specified operand, returning the same
7350 values as the libm <tt>sin</tt> functions would, and handles error conditions
7351 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7352
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7353</div>
7354
7355<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7356<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7357 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7358</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7359
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7360<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7361
7362<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7363<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
7364 floating point or vector of floating point type. Not all targets support all
7365 types however.</p>
7366
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7367<pre>
7368 declare float @llvm.cos.f32(float %Val)
7369 declare double @llvm.cos.f64(double %Val)
7370 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
7371 declare fp128 @llvm.cos.f128(fp128 %Val)
7372 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
7373</pre>
7374
7375<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7376<p>The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7377
7378<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7379<p>The argument and return value are floating point numbers of the same
7380 type.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7381
7382<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7383<p>This function returns the cosine of the specified operand, returning the same
7384 values as the libm <tt>cos</tt> functions would, and handles error conditions
7385 in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7386
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7387</div>
7388
7389<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7390<h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7391 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7392</h4>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7393
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7394<div>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7395
7396<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7397<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
7398 floating point or vector of floating point type. Not all targets support all
7399 types however.</p>
7400
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7401<pre>
7402 declare float @llvm.pow.f32(float %Val, float %Power)
7403 declare double @llvm.pow.f64(double %Val, double %Power)
7404 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
7405 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
7406 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
7407</pre>
7408
7409<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7410<p>The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
7411 specified (positive or negative) power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7412
7413<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7414<p>The second argument is a floating point power, and the first is a value to
7415 raise to that power.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7416
7417<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7418<p>This function returns the first value raised to the second power, returning
7419 the same values as the libm <tt>pow</tt> functions would, and handles error
7420 conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7421
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]7422</div>
7423
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7424<!-- _______________________________________________________________________ -->
7425<h4>
7426 <a name="int_exp">'<tt>llvm.exp.*</tt>' Intrinsic</a>
7427</h4>
7428
7429<div>
7430
7431<h5>Syntax:</h5>
7432<p>This is an overloaded intrinsic. You can use <tt>llvm.exp</tt> on any
7433 floating point or vector of floating point type. Not all targets support all
7434 types however.</p>
7435
7436<pre>
7437 declare float @llvm.exp.f32(float %Val)
7438 declare double @llvm.exp.f64(double %Val)
7439 declare x86_fp80 @llvm.exp.f80(x86_fp80 %Val)
7440 declare fp128 @llvm.exp.f128(fp128 %Val)
7441 declare ppc_fp128 @llvm.exp.ppcf128(ppc_fp128 %Val)
7442</pre>
7443
7444<h5>Overview:</h5>
7445<p>The '<tt>llvm.exp.*</tt>' intrinsics perform the exp function.</p>
7446
7447<h5>Arguments:</h5>
7448<p>The argument and return value are floating point numbers of the same
7449 type.</p>
7450
7451<h5>Semantics:</h5>
7452<p>This function returns the same values as the libm <tt>exp</tt> functions
7453 would, and handles error conditions in the same way.</p>
7454
7455</div>
7456
7457<!-- _______________________________________________________________________ -->
7458<h4>
7459 <a name="int_log">'<tt>llvm.log.*</tt>' Intrinsic</a>
7460</h4>
7461
7462<div>
7463
7464<h5>Syntax:</h5>
7465<p>This is an overloaded intrinsic. You can use <tt>llvm.log</tt> on any
7466 floating point or vector of floating point type. Not all targets support all
7467 types however.</p>
7468
7469<pre>
7470 declare float @llvm.log.f32(float %Val)
7471 declare double @llvm.log.f64(double %Val)
7472 declare x86_fp80 @llvm.log.f80(x86_fp80 %Val)
7473 declare fp128 @llvm.log.f128(fp128 %Val)
7474 declare ppc_fp128 @llvm.log.ppcf128(ppc_fp128 %Val)
7475</pre>
7476
7477<h5>Overview:</h5>
7478<p>The '<tt>llvm.log.*</tt>' intrinsics perform the log function.</p>
7479
7480<h5>Arguments:</h5>
7481<p>The argument and return value are floating point numbers of the same
7482 type.</p>
7483
7484<h5>Semantics:</h5>
7485<p>This function returns the same values as the libm <tt>log</tt> functions
7486 would, and handles error conditions in the same way.</p>
7487
Nick Lewycky1c929be2011-10-31 01:32:21 +0000[diff] [blame]7488</div>
7489
7490<!-- _______________________________________________________________________ -->
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7491<h4>
7492 <a name="int_fma">'<tt>llvm.fma.*</tt>' Intrinsic</a>
7493</h4>
7494
7495<div>
7496
7497<h5>Syntax:</h5>
7498<p>This is an overloaded intrinsic. You can use <tt>llvm.fma</tt> on any
7499 floating point or vector of floating point type. Not all targets support all
7500 types however.</p>
7501
7502<pre>
7503 declare float @llvm.fma.f32(float %a, float %b, float %c)
7504 declare double @llvm.fma.f64(double %a, double %b, double %c)
7505 declare x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
7506 declare fp128 @llvm.fma.f128(fp128 %a, fp128 %b, fp128 %c)
7507 declare ppc_fp128 @llvm.fma.ppcf128(ppc_fp128 %a, ppc_fp128 %b, ppc_fp128 %c)
7508</pre>
7509
7510<h5>Overview:</h5>
Cameron Zwarichabc43e62011-07-08 22:13:55 +0000[diff] [blame]7511<p>The '<tt>llvm.fma.*</tt>' intrinsics perform the fused multiply-add
Cameron Zwarich33390842011-07-08 21:39:21 +0000[diff] [blame]7512 operation.</p>
7513
7514<h5>Arguments:</h5>
7515<p>The argument and return value are floating point numbers of the same
7516 type.</p>
7517
7518<h5>Semantics:</h5>
7519<p>This function returns the same values as the libm <tt>fma</tt> functions
7520 would.</p>
7521
Dan Gohman4e9011c2011-05-23 21:13:03 +0000[diff] [blame]7522</div>
7523
Peter Collingbourne168a4c32012-07-03 12:25:40 +0000[diff] [blame]7524<!-- _______________________________________________________________________ -->
7525<h4>
7526 <a name="int_fabs">'<tt>llvm.fabs.*</tt>' Intrinsic</a>
7527</h4>
7528
7529<div>
7530
7531<h5>Syntax:</h5>
7532<p>This is an overloaded intrinsic. You can use <tt>llvm.fabs</tt> on any
7533 floating point or vector of floating point type. Not all targets support all
7534 types however.</p>
7535
7536<pre>
7537 declare float @llvm.fabs.f32(float %Val)
7538 declare double @llvm.fabs.f64(double %Val)
7539 declare x86_fp80 @llvm.fabs.f80(x86_fp80 %Val)
7540 declare fp128 @llvm.fabs.f128(fp128 %Val)
7541 declare ppc_fp128 @llvm.fabs.ppcf128(ppc_fp128 %Val)
7542</pre>
7543
7544<h5>Overview:</h5>
7545<p>The '<tt>llvm.fabs.*</tt>' intrinsics return the absolute value of
7546 the operand.</p>
7547
7548<h5>Arguments:</h5>
7549<p>The argument and return value are floating point numbers of the same
7550 type.</p>
7551
7552<h5>Semantics:</h5>
7553<p>This function returns the same values as the libm <tt>fabs</tt> functions
7554 would, and handles error conditions in the same way.</p>
7555
7556</div>
7557
Dan Gohman27db99f2012-07-26 17:43:27 +0000[diff] [blame]7558<!-- _______________________________________________________________________ -->
7559<h4>
7560 <a name="int_floor">'<tt>llvm.floor.*</tt>' Intrinsic</a>
7561</h4>
7562
7563<div>
7564
7565<h5>Syntax:</h5>
7566<p>This is an overloaded intrinsic. You can use <tt>llvm.floor</tt> on any
7567 floating point or vector of floating point type. Not all targets support all
7568 types however.</p>
7569
7570<pre>
7571 declare float @llvm.floor.f32(float %Val)
7572 declare double @llvm.floor.f64(double %Val)
7573 declare x86_fp80 @llvm.floor.f80(x86_fp80 %Val)
7574 declare fp128 @llvm.floor.f128(fp128 %Val)
7575 declare ppc_fp128 @llvm.floor.ppcf128(ppc_fp128 %Val)
7576</pre>
7577
7578<h5>Overview:</h5>
7579<p>The '<tt>llvm.floor.*</tt>' intrinsics return the floor of
7580 the operand.</p>
7581
7582<h5>Arguments:</h5>
7583<p>The argument and return value are floating point numbers of the same
7584 type.</p>
7585
7586<h5>Semantics:</h5>
7587<p>This function returns the same values as the libm <tt>floor</tt> functions
7588 would, and handles error conditions in the same way.</p>
7589
7590</div>
7591
NAKAMURA Takumi4b2e07a2011-10-31 13:04:26 +0000[diff] [blame]7592</div>
7593
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7594<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7595<h3>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7596 <a name="int_manip">Bit Manipulation Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7597</h3>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7598
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7599<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7600
7601<p>LLVM provides intrinsics for a few important bit manipulation operations.
7602 These allow efficient code generation for some algorithms.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7603
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7604<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7605<h4>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]7606 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7607</h4>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7608
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7609<div>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7610
7611<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7612<p>This is an overloaded intrinsic function. You can use bswap on any integer
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7613 type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
7614
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7615<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7616 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
7617 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
7618 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7619</pre>
7620
7621<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7622<p>The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
7623 values with an even number of bytes (positive multiple of 16 bits). These
7624 are useful for performing operations on data that is not in the target's
7625 native byte order.</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7626
7627<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7628<p>The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
7629 and low byte of the input i16 swapped. Similarly,
7630 the <tt>llvm.bswap.i32</tt> intrinsic returns an i32 value that has the four
7631 bytes of the input i32 swapped, so that if the input bytes are numbered 0, 1,
7632 2, 3 then the returned i32 will have its bytes in 3, 2, 1, 0 order.
7633 The <tt>llvm.bswap.i48</tt>, <tt>llvm.bswap.i64</tt> and other intrinsics
7634 extend this concept to additional even-byte lengths (6 bytes, 8 bytes and
7635 more, respectively).</p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]7636
7637</div>
7638
7639<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7640<h4>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]7641 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7642</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7643
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7644<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7645
7646<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]7647<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]7648 width, or on any vector with integer elements. Not all targets support all
7649 bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7650
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7651<pre>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7652 declare i8 @llvm.ctpop.i8(i8 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7653 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]7654 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]7655 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
7656 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7657 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]7658</pre>
7659
7660<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7661<p>The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set
7662 in a value.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7663
7664<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7665<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]7666 integer type, or a vector with integer elements.
7667 The return type must match the argument type.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7668
7669<h5>Semantics:</h5>
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7670<p>The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable, or within each
7671 element of a vector.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7672
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7673</div>
7674
7675<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7676<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7677 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7678</h4>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7679
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7680<div>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7681
7682<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7683<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]7684 integer bit width, or any vector whose elements are integers. Not all
7685 targets support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7686
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7687<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7688 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7689 declare i16 @llvm.ctlz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7690 declare i32 @llvm.ctlz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7691 declare i64 @llvm.ctlz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7692 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7693 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]7694</pre>
7695
7696<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7697<p>The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
7698 leading zeros in a variable.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7699
7700<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7701<p>The first argument is the value to be counted. This argument may be of any
7702 integer type, or a vectory with integer element type. The return type
7703 must match the first argument type.</p>
7704
7705<p>The second argument must be a constant and is a flag to indicate whether the
7706 intrinsic should ensure that a zero as the first argument produces a defined
7707 result. Historically some architectures did not provide a defined result for
7708 zero values as efficiently, and many algorithms are now predicated on
7709 avoiding zero-value inputs.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7710
7711<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7712<p>The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant)
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7713 zeros in a variable, or within each element of the vector.
7714 If <tt>src == 0</tt> then the result is the size in bits of the type of
7715 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7716 For example, <tt>llvm.ctlz(i32 2) = 30</tt>.</p>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7717
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]7718</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]7719
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7720<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7721<h4>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]7722 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7723</h4>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7724
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7725<div>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7726
7727<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7728<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]7729 integer bit width, or any vector of integer elements. Not all targets
7730 support all bit widths or vector types, however.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7731
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7732<pre>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7733 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7734 declare i16 @llvm.cttz.i16 (i16 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7735 declare i32 @llvm.cttz.i32 (i32 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7736 declare i64 @llvm.cttz.i64 (i64 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7737 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;, i1 &lt;is_zero_undef&gt;)
7738 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]7739</pre>
7740
7741<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7742<p>The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
7743 trailing zeros.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7744
7745<h5>Arguments:</h5>
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7746<p>The first argument is the value to be counted. This argument may be of any
7747 integer type, or a vectory with integer element type. The return type
7748 must match the first argument type.</p>
7749
7750<p>The second argument must be a constant and is a flag to indicate whether the
7751 intrinsic should ensure that a zero as the first argument produces a defined
7752 result. Historically some architectures did not provide a defined result for
7753 zero values as efficiently, and many algorithms are now predicated on
7754 avoiding zero-value inputs.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7755
7756<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7757<p>The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant)
Owen Andersonf1ac4652011-07-01 21:52:38 +0000[diff] [blame]7758 zeros in a variable, or within each element of a vector.
Chandler Carruth48b0bbf2011-12-12 04:36:04 +0000[diff] [blame]7759 If <tt>src == 0</tt> then the result is the size in bits of the type of
7760 <tt>src</tt> if <tt>is_zero_undef == 0</tt> and <tt>undef</tt> otherwise.
7761 For example, <tt>llvm.cttz(2) = 1</tt>.</p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7762
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]7763</div>
7764
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7765</div>
7766
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7767<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7768<h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7769 <a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7770</h3>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7771
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7772<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7773
7774<p>LLVM provides intrinsics for some arithmetic with overflow operations.</p>
Bill Wendlingda01af72009-02-08 04:04:40 +0000[diff] [blame]7775
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7776<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7777<h4>
7778 <a name="int_sadd_overflow">
7779 '<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics
7780 </a>
7781</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7782
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7783<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7784
7785<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7786<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]7787 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7788
7789<pre>
7790 declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
7791 declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7792 declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
7793</pre>
7794
7795<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7796<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7797 a signed addition of the two arguments, and indicate whether an overflow
7798 occurred during the signed summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7799
7800<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7801<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]7802 be of integer types of any bit width, but they must have the same bit
7803 width. The second element of the result structure must be of
7804 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7805 undergo signed addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7806
7807<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7808<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7809 a signed addition of the two variables. They return a structure &mdash; the
7810 first element of which is the signed summation, and the second element of
7811 which is a bit specifying if the signed summation resulted in an
7812 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7813
7814<h5>Examples:</h5>
7815<pre>
7816 %res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
7817 %sum = extractvalue {i32, i1} %res, 0
7818 %obit = extractvalue {i32, i1} %res, 1
7819 br i1 %obit, label %overflow, label %normal
7820</pre>
7821
7822</div>
7823
7824<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7825<h4>
7826 <a name="int_uadd_overflow">
7827 '<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics
7828 </a>
7829</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7830
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7831<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7832
7833<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7834<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]7835 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7836
7837<pre>
7838 declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
7839 declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7840 declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
7841</pre>
7842
7843<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7844<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7845 an unsigned addition of the two arguments, and indicate whether a carry
7846 occurred during the unsigned summation.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7847
7848<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7849<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]7850 be of integer types of any bit width, but they must have the same bit
7851 width. The second element of the result structure must be of
7852 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7853 undergo unsigned addition.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7854
7855<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7856<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7857 an unsigned addition of the two arguments. They return a structure &mdash;
7858 the first element of which is the sum, and the second element of which is a
7859 bit specifying if the unsigned summation resulted in a carry.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7860
7861<h5>Examples:</h5>
7862<pre>
7863 %res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
7864 %sum = extractvalue {i32, i1} %res, 0
7865 %obit = extractvalue {i32, i1} %res, 1
7866 br i1 %obit, label %carry, label %normal
7867</pre>
7868
7869</div>
7870
7871<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7872<h4>
7873 <a name="int_ssub_overflow">
7874 '<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics
7875 </a>
7876</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7877
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7878<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7879
7880<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7881<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]7882 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7883
7884<pre>
7885 declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
7886 declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7887 declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
7888</pre>
7889
7890<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7891<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7892 a signed subtraction of the two arguments, and indicate whether an overflow
7893 occurred during the signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7894
7895<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7896<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]7897 be of integer types of any bit width, but they must have the same bit
7898 width. The second element of the result structure must be of
7899 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7900 undergo signed subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7901
7902<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7903<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7904 a signed subtraction of the two arguments. They return a structure &mdash;
7905 the first element of which is the subtraction, and the second element of
7906 which is a bit specifying if the signed subtraction resulted in an
7907 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7908
7909<h5>Examples:</h5>
7910<pre>
7911 %res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
7912 %sum = extractvalue {i32, i1} %res, 0
7913 %obit = extractvalue {i32, i1} %res, 1
7914 br i1 %obit, label %overflow, label %normal
7915</pre>
7916
7917</div>
7918
7919<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7920<h4>
7921 <a name="int_usub_overflow">
7922 '<tt>llvm.usub.with.overflow.*</tt>' Intrinsics
7923 </a>
7924</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7925
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7926<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7927
7928<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7929<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]7930 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7931
7932<pre>
7933 declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
7934 declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7935 declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
7936</pre>
7937
7938<h5>Overview:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7939<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7940 an unsigned subtraction of the two arguments, and indicate whether an
7941 overflow occurred during the unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7942
7943<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7944<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]7945 be of integer types of any bit width, but they must have the same bit
7946 width. The second element of the result structure must be of
7947 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7948 undergo unsigned subtraction.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7949
7950<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7951<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7952 an unsigned subtraction of the two arguments. They return a structure &mdash;
7953 the first element of which is the subtraction, and the second element of
7954 which is a bit specifying if the unsigned subtraction resulted in an
7955 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7956
7957<h5>Examples:</h5>
7958<pre>
7959 %res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
7960 %sum = extractvalue {i32, i1} %res, 0
7961 %obit = extractvalue {i32, i1} %res, 1
7962 br i1 %obit, label %overflow, label %normal
7963</pre>
7964
7965</div>
7966
7967<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]7968<h4>
7969 <a name="int_smul_overflow">
7970 '<tt>llvm.smul.with.overflow.*</tt>' Intrinsics
7971 </a>
7972</h4>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7973
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]7974<div>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7975
7976<h5>Syntax:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7977<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]7978 on any integer bit width.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7979
7980<pre>
7981 declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
7982 declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
7983 declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
7984</pre>
7985
7986<h5>Overview:</h5>
7987
7988<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]7989 a signed multiplication of the two arguments, and indicate whether an
7990 overflow occurred during the signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7991
7992<h5>Arguments:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7993<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]7994 be of integer types of any bit width, but they must have the same bit
7995 width. The second element of the result structure must be of
7996 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
7997 undergo signed multiplication.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]7998
7999<h5>Semantics:</h5>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]8000<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8001 a signed multiplication of the two arguments. They return a structure &mdash;
8002 the first element of which is the multiplication, and the second element of
8003 which is a bit specifying if the signed multiplication resulted in an
8004 overflow.</p>
Bill Wendlingac1df8e2009-02-08 01:40:31 +0000[diff] [blame]8005
8006<h5>Examples:</h5>
8007<pre>
8008 %res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
8009 %sum = extractvalue {i32, i1} %res, 0
8010 %obit = extractvalue {i32, i1} %res, 1
8011 br i1 %obit, label %overflow, label %normal
8012</pre>
8013
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]8014</div>
8015
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8016<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8017<h4>
8018 <a name="int_umul_overflow">
8019 '<tt>llvm.umul.with.overflow.*</tt>' Intrinsics
8020 </a>
8021</h4>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8022
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8023<div>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8024
8025<h5>Syntax:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8026<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]8027 on any integer bit width.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8028
8029<pre>
8030 declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
8031 declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8032 declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
8033</pre>
8034
8035<h5>Overview:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8036<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8037 a unsigned multiplication of the two arguments, and indicate whether an
8038 overflow occurred during the unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8039
8040<h5>Arguments:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8041<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]8042 be of integer types of any bit width, but they must have the same bit
8043 width. The second element of the result structure must be of
8044 type <tt>i1</tt>. <tt>%a</tt> and <tt>%b</tt> are the two values that will
8045 undergo unsigned multiplication.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8046
8047<h5>Semantics:</h5>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8048<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8049 an unsigned multiplication of the two arguments. They return a structure
8050 &mdash; the first element of which is the multiplication, and the second
8051 element of which is a bit specifying if the unsigned multiplication resulted
8052 in an overflow.</p>
Bill Wendling41b485c2009-02-08 23:00:09 +0000[diff] [blame]8053
8054<h5>Examples:</h5>
8055<pre>
8056 %res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
8057 %sum = extractvalue {i32, i1} %res, 0
8058 %obit = extractvalue {i32, i1} %res, 1
8059 br i1 %obit, label %overflow, label %normal
8060</pre>
8061
8062</div>
8063
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8064</div>
8065
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8066<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8067<h3>
Lang Hames5afba6f2012-06-05 19:07:46 +0000[diff] [blame]8068 <a name="spec_arithmetic">Specialised Arithmetic Intrinsics</a>
8069</h3>
8070
8071<!-- _______________________________________________________________________ -->
8072
8073<h4>
8074 <a name="fmuladd">'<tt>llvm.fmuladd.*</tt>' Intrinsic</a>
8075</h4>
8076
8077<div>
8078
8079<h5>Syntax:</h5>
8080<pre>
8081 declare float @llvm.fmuladd.f32(float %a, float %b, float %c)
8082 declare double @llvm.fmuladd.f64(double %a, double %b, double %c)
8083</pre>
8084
8085<h5>Overview:</h5>
8086<p>The '<tt>llvm.fmuladd.*</tt>' intrinsic functions represent multiply-add
8087expressions that can be fused if the code generator determines that the fused
8088expression would be legal and efficient.</p>
8089
8090<h5>Arguments:</h5>
8091<p>The '<tt>llvm.fmuladd.*</tt>' intrinsics each take three arguments: two
8092multiplicands, a and b, and an addend c.</p>
8093
8094<h5>Semantics:</h5>
8095<p>The expression:</p>
8096<pre>
8097 %0 = call float @llvm.fmuladd.f32(%a, %b, %c)
8098</pre>
8099<p>is equivalent to the expression a * b + c, except that rounding will not be
8100performed between the multiplication and addition steps if the code generator
8101fuses the operations. Fusion is not guaranteed, even if the target platform
8102supports it. If a fused multiply-add is required the corresponding llvm.fma.*
8103intrinsic function should be used instead.</p>
8104
8105<h5>Examples:</h5>
8106<pre>
8107 %r2 = call float @llvm.fmuladd.f32(float %a, float %b, float %c) ; yields {float}:r2 = (a * b) + c
8108</pre>
8109
8110</div>
8111
8112<!-- ======================================================================= -->
8113<h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8114 <a name="int_fp16">Half Precision Floating Point Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8115</h3>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8116
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8117<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8118
Tobias Grosser057beb82012-05-24 15:59:06 +0000[diff] [blame]8119<p>For most target platforms, half precision floating point is a storage-only
8120 format. This means that it is
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]8121 a dense encoding (in memory) but does not support computation in the
8122 format.</p>
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]8123
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]8124<p>This means that code must first load the half-precision floating point
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]8125 value as an i16, then convert it to float with <a
8126 href="#int_convert_from_fp16"><tt>llvm.convert.from.fp16</tt></a>.
8127 Computation can then be performed on the float value (including extending to
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]8128 double etc). To store the value back to memory, it is first converted to
8129 float if needed, then converted to i16 with
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]8130 <a href="#int_convert_to_fp16"><tt>llvm.convert.to.fp16</tt></a>, then
8131 storing as an i16 value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8132
8133<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8134<h4>
8135 <a name="int_convert_to_fp16">
8136 '<tt>llvm.convert.to.fp16</tt>' Intrinsic
8137 </a>
8138</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8139
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8140<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8141
8142<h5>Syntax:</h5>
8143<pre>
8144 declare i16 @llvm.convert.to.fp16(f32 %a)
8145</pre>
8146
8147<h5>Overview:</h5>
8148<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
8149 a conversion from single precision floating point format to half precision
8150 floating point format.</p>
8151
8152<h5>Arguments:</h5>
8153<p>The intrinsic function contains single argument - the value to be
8154 converted.</p>
8155
8156<h5>Semantics:</h5>
8157<p>The '<tt>llvm.convert.to.fp16</tt>' intrinsic function performs
8158 a conversion from single precision floating point format to half precision
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]8159 floating point format. The return value is an <tt>i16</tt> which
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]8160 contains the converted number.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8161
8162<h5>Examples:</h5>
8163<pre>
8164 %res = call i16 @llvm.convert.to.fp16(f32 %a)
8165 store i16 %res, i16* @x, align 2
8166</pre>
8167
8168</div>
8169
8170<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8171<h4>
8172 <a name="int_convert_from_fp16">
8173 '<tt>llvm.convert.from.fp16</tt>' Intrinsic
8174 </a>
8175</h4>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8176
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8177<div>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8178
8179<h5>Syntax:</h5>
8180<pre>
8181 declare f32 @llvm.convert.from.fp16(i16 %a)
8182</pre>
8183
8184<h5>Overview:</h5>
8185<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs
8186 a conversion from half precision floating point format to single precision
8187 floating point format.</p>
8188
8189<h5>Arguments:</h5>
8190<p>The intrinsic function contains single argument - the value to be
8191 converted.</p>
8192
8193<h5>Semantics:</h5>
8194<p>The '<tt>llvm.convert.from.fp16</tt>' intrinsic function performs a
Chris Lattner0cec9c82010-03-15 04:12:21 +0000[diff] [blame]8195 conversion from half single precision floating point format to single
Chris Lattner82c3dc62010-03-14 23:03:31 +0000[diff] [blame]8196 precision floating point format. The input half-float value is represented by
8197 an <tt>i16</tt> value.</p>
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8198
8199<h5>Examples:</h5>
8200<pre>
8201 %a = load i16* @x, align 2
8202 %res = call f32 @llvm.convert.from.fp16(i16 %a)
8203</pre>
8204
8205</div>
8206
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8207</div>
8208
Anton Korobeynikovf02e7302010-03-14 18:42:47 +0000[diff] [blame]8209<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8210<h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8211 <a name="int_debugger">Debugger Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8212</h3>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8213
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8214<div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8215
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8216<p>The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt>
8217 prefix), are described in
8218 the <a href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source
8219 Level Debugging</a> document.</p>
8220
8221</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]8222
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8223<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8224<h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8225 <a name="int_eh">Exception Handling Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8226</h3>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8227
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8228<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8229
8230<p>The LLVM exception handling intrinsics (which all start with
8231 <tt>llvm.eh.</tt> prefix), are described in
8232 the <a href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
8233 Handling</a> document.</p>
8234
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8235</div>
8236
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8237<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8238<h3>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8239 <a name="int_trampoline">Trampoline Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8240</h3>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8241
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8242<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8243
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8244<p>These intrinsics make it possible to excise one parameter, marked with
Dan Gohmanff235352010-07-02 23:18:08 +0000[diff] [blame]8245 the <a href="#nest"><tt>nest</tt></a> attribute, from a function.
8246 The result is a callable
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8247 function pointer lacking the nest parameter - the caller does not need to
8248 provide a value for it. Instead, the value to use is stored in advance in a
8249 "trampoline", a block of memory usually allocated on the stack, which also
8250 contains code to splice the nest value into the argument list. This is used
8251 to implement the GCC nested function address extension.</p>
8252
8253<p>For example, if the function is
8254 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
8255 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as
8256 follows:</p>
8257
Benjamin Kramer26fe25f2010-07-13 12:26:09 +0000[diff] [blame]8258<pre class="doc_code">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]8259 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
8260 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8261 call i8* @llvm.init.trampoline(i8* %tramp1, i8* bitcast (i32 (i8*, i32, i32)* @f to i8*), i8* %nval)
8262 %p = call i8* @llvm.adjust.trampoline(i8* %tramp1)
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]8263 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8264</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8265
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8266<p>The call <tt>%val = call i32 %fp(i32 %x, i32 %y)</tt> is then equivalent
8267 to <tt>%val = call i32 %f(i8* %nval, i32 %x, i32 %y)</tt>.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8268
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8269<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8270<h4>
8271 <a name="int_it">
8272 '<tt>llvm.init.trampoline</tt>' Intrinsic
8273 </a>
8274</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8275
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8276<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8277
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8278<h5>Syntax:</h5>
8279<pre>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8280 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]8281</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8282
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8283<h5>Overview:</h5>
Duncan Sands4a544a72011-09-06 13:37:06 +0000[diff] [blame]8284<p>This fills the memory pointed to by <tt>tramp</tt> with executable code,
8285 turning it into a trampoline.</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8286
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8287<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8288<p>The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
8289 pointers. The <tt>tramp</tt> argument must point to a sufficiently large and
8290 sufficiently aligned block of memory; this memory is written to by the
8291 intrinsic. Note that the size and the alignment are target-specific - LLVM
8292 currently provides no portable way of determining them, so a front-end that
8293 generates this intrinsic needs to have some target-specific knowledge.
8294 The <tt>func</tt> argument must hold a function bitcast to
8295 an <tt>i8*</tt>.</p>
8296
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8297<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8298<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]8299 dependent code, turning it into a function. Then <tt>tramp</tt> needs to be
8300 passed to <a href="#int_at">llvm.adjust.trampoline</a> to get a pointer
8301 which can be <a href="#int_trampoline">bitcast (to a new function) and
8302 called</a>. The new function's signature is the same as that of
8303 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
8304 removed. At most one such <tt>nest</tt> argument is allowed, and it must be of
8305 pointer type. Calling the new function is equivalent to calling <tt>func</tt>
8306 with the same argument list, but with <tt>nval</tt> used for the missing
8307 <tt>nest</tt> argument. If, after calling <tt>llvm.init.trampoline</tt>, the
8308 memory pointed to by <tt>tramp</tt> is modified, then the effect of any later call
8309 to the returned function pointer is undefined.</p>
8310</div>
8311
8312<!-- _______________________________________________________________________ -->
8313<h4>
8314 <a name="int_at">
8315 '<tt>llvm.adjust.trampoline</tt>' Intrinsic
8316 </a>
8317</h4>
8318
8319<div>
8320
8321<h5>Syntax:</h5>
8322<pre>
8323 declare i8* @llvm.adjust.trampoline(i8* &lt;tramp&gt;)
8324</pre>
8325
8326<h5>Overview:</h5>
8327<p>This performs any required machine-specific adjustment to the address of a
8328 trampoline (passed as <tt>tramp</tt>).</p>
8329
8330<h5>Arguments:</h5>
8331<p><tt>tramp</tt> must point to a block of memory which already has trampoline code
8332 filled in by a previous call to <a href="#int_it"><tt>llvm.init.trampoline</tt>
8333 </a>.</p>
8334
8335<h5>Semantics:</h5>
8336<p>On some architectures the address of the code to be executed needs to be
8337 different to the address where the trampoline is actually stored. This
8338 intrinsic returns the executable address corresponding to <tt>tramp</tt>
8339 after performing the required machine specific adjustments.
8340 The pointer returned can then be <a href="#int_trampoline"> bitcast and
8341 executed</a>.
8342</p>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8343
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8344</div>
8345
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8346</div>
8347
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]8348<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8349<h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8350 <a name="int_memorymarkers">Memory Use Markers</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8351</h3>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8352
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8353<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8354
8355<p>This class of intrinsics exists to information about the lifetime of memory
8356 objects and ranges where variables are immutable.</p>
8357
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8358<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8359<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8360 <a name="int_lifetime_start">'<tt>llvm.lifetime.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8361</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8362
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8363<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8364
8365<h5>Syntax:</h5>
8366<pre>
8367 declare void @llvm.lifetime.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8368</pre>
8369
8370<h5>Overview:</h5>
8371<p>The '<tt>llvm.lifetime.start</tt>' intrinsic specifies the start of a memory
8372 object's lifetime.</p>
8373
8374<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8375<p>The first argument is a constant integer representing the size of the
8376 object, or -1 if it is variable sized. The second argument is a pointer to
8377 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8378
8379<h5>Semantics:</h5>
8380<p>This intrinsic indicates that before this point in the code, the value of the
8381 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]8382 never be used and has an undefined value. A load from the pointer that
8383 precedes this intrinsic can be replaced with
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8384 <tt>'<a href="#undefvalues">undef</a>'</tt>.</p>
8385
8386</div>
8387
8388<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8389<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8390 <a name="int_lifetime_end">'<tt>llvm.lifetime.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8391</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8392
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8393<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8394
8395<h5>Syntax:</h5>
8396<pre>
8397 declare void @llvm.lifetime.end(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8398</pre>
8399
8400<h5>Overview:</h5>
8401<p>The '<tt>llvm.lifetime.end</tt>' intrinsic specifies the end of a memory
8402 object's lifetime.</p>
8403
8404<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8405<p>The first argument is a constant integer representing the size of the
8406 object, or -1 if it is variable sized. The second argument is a pointer to
8407 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8408
8409<h5>Semantics:</h5>
8410<p>This intrinsic indicates that after this point in the code, the value of the
8411 memory pointed to by <tt>ptr</tt> is dead. This means that it is known to
8412 never be used and has an undefined value. Any stores into the memory object
8413 following this intrinsic may be removed as dead.
8414
8415</div>
8416
8417<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8418<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8419 <a name="int_invariant_start">'<tt>llvm.invariant.start</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8420</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8421
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8422<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8423
8424<h5>Syntax:</h5>
8425<pre>
Nick Lewycky29b6cb42010-11-30 04:13:41 +0000[diff] [blame]8426 declare {}* @llvm.invariant.start(i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8427</pre>
8428
8429<h5>Overview:</h5>
8430<p>The '<tt>llvm.invariant.start</tt>' intrinsic specifies that the contents of
8431 a memory object will not change.</p>
8432
8433<h5>Arguments:</h5>
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8434<p>The first argument is a constant integer representing the size of the
8435 object, or -1 if it is variable sized. The second argument is a pointer to
8436 the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8437
8438<h5>Semantics:</h5>
8439<p>This intrinsic indicates that until an <tt>llvm.invariant.end</tt> that uses
8440 the return value, the referenced memory location is constant and
8441 unchanging.</p>
8442
8443</div>
8444
8445<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8446<h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8447 <a name="int_invariant_end">'<tt>llvm.invariant.end</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8448</h4>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8449
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8450<div>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8451
8452<h5>Syntax:</h5>
8453<pre>
8454 declare void @llvm.invariant.end({}* &lt;start&gt;, i64 &lt;size&gt;, i8* nocapture &lt;ptr&gt;)
8455</pre>
8456
8457<h5>Overview:</h5>
8458<p>The '<tt>llvm.invariant.end</tt>' intrinsic specifies that the contents of
8459 a memory object are mutable.</p>
8460
8461<h5>Arguments:</h5>
8462<p>The first argument is the matching <tt>llvm.invariant.start</tt> intrinsic.
Nick Lewycky321333e2009-10-13 07:57:33 +0000[diff] [blame]8463 The second argument is a constant integer representing the size of the
8464 object, or -1 if it is variable sized and the third argument is a pointer
8465 to the object.</p>
Nick Lewyckycc271862009-10-13 07:03:23 +0000[diff] [blame]8466
8467<h5>Semantics:</h5>
8468<p>This intrinsic indicates that the memory is mutable again.</p>
8469
8470</div>
8471
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8472</div>
8473
Andrew Lenharth22c5c1b2008-02-16 01:24:58 +0000[diff] [blame]8474<!-- ======================================================================= -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8475<h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8476 <a name="int_general">General Intrinsics</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8477</h3>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8478
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8479<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8480
8481<p>This class of intrinsics is designed to be generic and has no specific
8482 purpose.</p>
8483
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8484<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8485<h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8486 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8487</h4>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8488
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8489<div>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8490
8491<h5>Syntax:</h5>
8492<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8493 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]8494</pre>
8495
8496<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8497<p>The '<tt>llvm.var.annotation</tt>' intrinsic.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8498
8499<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8500<p>The first argument is a pointer to a value, the second is a pointer to a
8501 global string, the third is a pointer to a global string which is the source
8502 file name, and the last argument is the line number.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8503
8504<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8505<p>This intrinsic allows annotation of local variables with arbitrary strings.
8506 This can be useful for special purpose optimizations that want to look for
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8507 these annotations. These have no other defined use; they are ignored by code
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8508 generation and optimization.</p>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8509
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]8510</div>
8511
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8512<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8513<h4>
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]8514 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8515</h4>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8516
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8517<div>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8518
8519<h5>Syntax:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8520<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
8521 any integer bit width.</p>
8522
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8523<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8524 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8525 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8526 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8527 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt;)
8528 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]8529</pre>
8530
8531<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8532<p>The '<tt>llvm.annotation</tt>' intrinsic.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8533
8534<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8535<p>The first argument is an integer value (result of some expression), the
8536 second is a pointer to a global string, the third is a pointer to a global
8537 string which is the source file name, and the last argument is the line
8538 number. It returns the value of the first argument.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8539
8540<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8541<p>This intrinsic allows annotations to be put on arbitrary expressions with
8542 arbitrary strings. This can be useful for special purpose optimizations that
John Criswelle865c032011-08-19 16:57:55 +0000[diff] [blame]8543 want to look for these annotations. These have no other defined use; they
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8544 are ignored by code generation and optimization.</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8545
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]8546</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]8547
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8548<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8549<h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8550 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8551</h4>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8552
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8553<div>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8554
8555<h5>Syntax:</h5>
8556<pre>
Chris Lattner86208902012-05-27 23:20:41 +0000[diff] [blame]8557 declare void @llvm.trap() noreturn nounwind
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8558</pre>
8559
8560<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8561<p>The '<tt>llvm.trap</tt>' intrinsic.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8562
8563<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8564<p>None.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8565
8566<h5>Semantics:</h5>
John Criswell99534f32012-05-16 00:26:51 +0000[diff] [blame]8567<p>This intrinsic is lowered to the target dependent trap instruction. If the
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8568 target does not have a trap instruction, this intrinsic will be lowered to
John Criswell99534f32012-05-16 00:26:51 +0000[diff] [blame]8569 a call of the <tt>abort()</tt> function.</p>
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8570
Anton Korobeynikov4cb86182008-01-15 22:31:34 +0000[diff] [blame]8571</div>
8572
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8573<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8574<h4>
Dan Gohmana6063c62012-05-14 18:58:10 +0000[diff] [blame]8575 <a name="int_debugtrap">'<tt>llvm.debugtrap</tt>' Intrinsic</a>
Dan Gohmand4347e12012-05-11 00:19:32 +0000[diff] [blame]8576</h4>
8577
8578<div>
8579
8580<h5>Syntax:</h5>
8581<pre>
Chris Lattner86208902012-05-27 23:20:41 +0000[diff] [blame]8582 declare void @llvm.debugtrap() nounwind
Dan Gohmand4347e12012-05-11 00:19:32 +0000[diff] [blame]8583</pre>
8584
8585<h5>Overview:</h5>
Dan Gohmana6063c62012-05-14 18:58:10 +0000[diff] [blame]8586<p>The '<tt>llvm.debugtrap</tt>' intrinsic.</p>
Dan Gohmand4347e12012-05-11 00:19:32 +0000[diff] [blame]8587
8588<h5>Arguments:</h5>
8589<p>None.</p>
8590
8591<h5>Semantics:</h5>
8592<p>This intrinsic is lowered to code which is intended to cause an execution
8593 trap with the intention of requesting the attention of a debugger.</p>
8594
8595</div>
8596
8597<!-- _______________________________________________________________________ -->
8598<h4>
Misha Brukmandccb0252008-11-22 23:55:29 +0000[diff] [blame]8599 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8600</h4>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8601
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8602<div>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8603
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8604<h5>Syntax:</h5>
8605<pre>
Dan Gohman3dfb3cf2010-05-28 17:07:41 +0000[diff] [blame]8606 declare void @llvm.stackprotector(i8* &lt;guard&gt;, i8** &lt;slot&gt;)
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8607</pre>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8608
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8609<h5>Overview:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8610<p>The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and
8611 stores it onto the stack at <tt>slot</tt>. The stack slot is adjusted to
8612 ensure that it is placed on the stack before local variables.</p>
8613
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8614<h5>Arguments:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8615<p>The <tt>llvm.stackprotector</tt> intrinsic requires two pointer
8616 arguments. The first argument is the value loaded from the stack
8617 guard <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt>
8618 that has enough space to hold the value of the guard.</p>
8619
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8620<h5>Semantics:</h5>
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8621<p>This intrinsic causes the prologue/epilogue inserter to force the position of
8622 the <tt>AllocaInst</tt> stack slot to be before local variables on the
8623 stack. This is to ensure that if a local variable on the stack is
8624 overwritten, it will destroy the value of the guard. When the function exits,
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8625 the guard on the stack is checked against the original guard. If they are
Bill Wendlinge910b4c2009-07-20 02:29:24 +0000[diff] [blame]8626 different, then the program aborts by calling the <tt>__stack_chk_fail()</tt>
8627 function.</p>
8628
Bill Wendling69e4adb2008-11-19 05:56:17 +0000[diff] [blame]8629</div>
8630
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8631<!-- _______________________________________________________________________ -->
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8632<h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8633 <a name="int_objectsize">'<tt>llvm.objectsize</tt>' Intrinsic</a>
NAKAMURA Takumi05d02652011-04-18 23:59:50 +0000[diff] [blame]8634</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8635
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8636<div>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8637
8638<h5>Syntax:</h5>
8639<pre>
Nuno Lopes23e75da2012-05-22 15:25:31 +0000[diff] [blame]8640 declare i32 @llvm.objectsize.i32(i8* &lt;object&gt;, i1 &lt;min&gt;)
8641 declare i64 @llvm.objectsize.i64(i8* &lt;object&gt;, i1 &lt;min&gt;)
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8642</pre>
8643
8644<h5>Overview:</h5>
Bill Wendling1b383ba2010-10-27 01:07:41 +0000[diff] [blame]8645<p>The <tt>llvm.objectsize</tt> intrinsic is designed to provide information to
8646 the optimizers to determine at compile time whether a) an operation (like
8647 memcpy) will overflow a buffer that corresponds to an object, or b) that a
8648 runtime check for overflow isn't necessary. An object in this context means
8649 an allocation of a specific class, structure, array, or other object.</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8650
8651<h5>Arguments:</h5>
Nuno Lopes23e75da2012-05-22 15:25:31 +0000[diff] [blame]8652<p>The <tt>llvm.objectsize</tt> intrinsic takes two arguments. The first
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8653 argument is a pointer to or into the <tt>object</tt>. The second argument
Nuno Lopes23e75da2012-05-22 15:25:31 +0000[diff] [blame]8654 is a boolean and determines whether <tt>llvm.objectsize</tt> returns 0 (if
8655 true) or -1 (if false) when the object size is unknown.
8656 The second argument only accepts constants.</p>
Eric Christopher8295a0a2009-12-23 00:29:49 +0000[diff] [blame]8657
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8658<h5>Semantics:</h5>
Nuno Lopes30759542012-05-09 15:52:43 +0000[diff] [blame]8659<p>The <tt>llvm.objectsize</tt> intrinsic is lowered to a constant representing
8660 the size of the object concerned. If the size cannot be determined at compile
Nuno Lopes23e75da2012-05-22 15:25:31 +0000[diff] [blame]8661 time, <tt>llvm.objectsize</tt> returns <tt>i32/i64 -1 or 0</tt>
8662 (depending on the <tt>min</tt> argument).</p>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8663
8664</div>
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8665<!-- _______________________________________________________________________ -->
8666<h4>
8667 <a name="int_expect">'<tt>llvm.expect</tt>' Intrinsic</a>
8668</h4>
Eric Christopher0e671492009-11-30 08:03:53 +0000[diff] [blame]8669
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8670<div>
8671
8672<h5>Syntax:</h5>
8673<pre>
8674 declare i32 @llvm.expect.i32(i32 &lt;val&gt;, i32 &lt;expected_val&gt;)
8675 declare i64 @llvm.expect.i64(i64 &lt;val&gt;, i64 &lt;expected_val&gt;)
8676</pre>
8677
8678<h5>Overview:</h5>
8679<p>The <tt>llvm.expect</tt> intrinsic provides information about expected (the
8680 most probable) value of <tt>val</tt>, which can be used by optimizers.</p>
8681
8682<h5>Arguments:</h5>
8683<p>The <tt>llvm.expect</tt> intrinsic takes two arguments. The first
8684 argument is a value. The second argument is an expected value, this needs to
8685 be a constant value, variables are not allowed.</p>
8686
8687<h5>Semantics:</h5>
8688<p>This intrinsic is lowered to the <tt>val</tt>.</p>
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8689</div>
8690
Nuno Lopesb0c76d72012-07-05 17:37:07 +0000[diff] [blame]8691<!-- _______________________________________________________________________ -->
8692<h4>
8693 <a name="int_donothing">'<tt>llvm.donothing</tt>' Intrinsic</a>
8694</h4>
8695
8696<div>
8697
8698<h5>Syntax:</h5>
8699<pre>
8700 declare void @llvm.donothing() nounwind readnone
8701</pre>
8702
8703<h5>Overview:</h5>
8704<p>The <tt>llvm.donothing</tt> intrinsic doesn't perform any operation. It's the
8705only intrinsic that can be called with an invoke instruction.</p>
8706
8707<h5>Arguments:</h5>
8708<p>None.</p>
8709
8710<h5>Semantics:</h5>
8711<p>This intrinsic does nothing, and it's removed by optimizers and ignored by
8712codegen.</p>
8713</div>
8714
NAKAMURA Takumif5af6ad2011-04-23 00:30:22 +0000[diff] [blame]8715</div>
8716
Jakub Staszakb170e2d2011-12-04 18:29:26 +0000[diff] [blame]8717</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8718<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]8719<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8720<address>
8721 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8722 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8723 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]8724 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]8725
8726 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
NAKAMURA Takumib9a33632011-04-09 02:13:37 +0000[diff] [blame]8727 <a href="http://llvm.org/">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]8728 Last modified: $Date$
8729</address>
Chris Lattnerc7d3ab32008-01-04 04:33:49 +0000[diff] [blame]8730
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]8731</body>
8732</html>